Home

Series 8000 Xpress PSE User Guide

1. Table B 3 Restarting Cause Codes X890 304751 Issue 1 B 3 Rev 0 B 4 X 25 X 75 Diagnostic Codes The diagnostic codes may be present in clearing reset and restart packets ey 1 2 10 1 2 16 X890 304751 Issue 1 NO ADDITIONAL INFORMATION Invalid P S Invalid P R PACKET TYPE INVALID Packet Type Invalid for State R1 Packet Type Invalid for State R2 Packet Type Invalid for State R3 Packet Type Invalid for State P1 Packet Type Invalid for State P2 Packet Type Invalid for State P3 Packet Type Invalid for State P4 Packet Type Invalid for State P6 Packet Type Invalid for State P7 Packet Type Invalid for State D1 Packet Type Invalid for State D2 Packet Type Invalid for State D3 PACKET NOT ALLOWED Unidentifiable Packet Call on one way logical channel Invalid Packet Type on PVC Packet on unassigned logical channel Reject not subscribed to Packet too short Packet too long Invalid general format identifier Table B 4 Diagnostic Codes B 4 Rev 0 Restart packet with non zero in bits 5 to 16 Packet type not compatible with facility Unauthorised Interrupt Confirmation Unauthorised Interrupt TIMER EXPIRED Timer expired for incoming call Timer expired for clear indication Timer expired for reset indication Timer expired for restart indication CALL SET UP CALL CLEARING PROBLEM Facility code not allowed Facility parameter not allowed valig cala address No logical channels available for this cal
2. Case Communications Limited has made all reasonable efforts to ensure the accuracy of the content of this document but the information contained herein does not constitute a warranty of performance of the equipment and or software described and no specifications given form part of any contract This document does not constitute a licence to use or copy any software described herein and any such software must only be used in accordance with the terms of the licence supplied therewith Case Communications Limited reserves the right to make alterations to the equipment and software described without notice and assumes no liability for any loss or damage caused as a result of use of this document whether because of out of date or inaccurate information or otherwise Product and manufacturers names referred to in this document are used for identification purposes only and Case Communications Limited acknowledges the intellectual property rights of their respective owners in the same This document is the copyright of Case Communications Limited and may not be reproduced copied or stored in any computerised retrieval system by any means whatsoever without the express written permission of Case Communications Limited Published by Case Communications Technical Publications Department X890 304751 Issue 1 0 2 Rev 1 Preface A short description of each of the Xpress PSEs is given in the Family Overview part number X890 304351 This User Guide des
3. Applications which reside on an ACM may use their own operating system and need not be written in the C programming language Co resident applications must run under the Xpress operating system be written in C and generally be well behaved Unless otherwise specifically stated the remainder of this appendix is concerned only with imported applications X890 304751 Issue 1 J 1 Rev 0 J 2 Network Architecture Within the context of the Xpress network architecture an application appears as an X 25 DTE Applications are addressed with Xpress network addresses Applications may be managed as network elements by the Cray 5X50 NMC Applications interface to the Xpress network via UNIX like sockets Xpress provides the following four network services J 2 1 X 25 Applications may originate and receive X 25 calls to from other applications or X 25 DTEs Xpress provides resilient re routing of calls and protects user data against internal network failures Xpress supports X 25 1980 84 88 Xpress also supports the ISO CONS see ISO 8878 except for receipt confirmation and not routing on ISO NSAP addresses it carries NSAP addresses transparently Future versions of the Xpress may allow applications to use PVCs The X 25 service is accessed using the AF_X25 SOCK_STREAM type of socket J 2 2 Network Management Service Applications within an Xpress network may interact with the Cray 5X50 NMC They communicate d
4. Incoming Called Calling address translation ICAT is performed as a Call Request packet enters the PSE at an X 25 X 75 port Outgoing Called Calling Address translation OCAT is performed as a Call Request packet leaves the PSE at the port Each translation is driven by a user configurable table The table is set up using the Routing Specification Incoming or Outgoing Called Calling Address Translation Source calling Address Translation edit table for port SCreen 4 5 3 Incoming Called Calling Address Translation ICAT Incoming address translation is performed when a call first enters the PSE at an X 25 port At each port there are two translation tables One table specifies the translations performed on called addresses the other table specifies translations on calling addresses Each table is organised as an ordered list of match address and substitute address pairs Call Request packets received at the port have their addresses compared against each table entry in sequence until a match is found If a match is found the address in the packet is replaced by the corresponding substitute address An address which does not match any entry in the appropriate table is allowed through unmodified Valid characters in the table addresses are the digits 0 to 9 Space characters in addresses are ignored A special wildcard character n allows any individual digit to be matched In the substitute address a wildcard character takes the val
5. Figure A 1 ISO 7 Layer Model for Open Systems Interconnection The International Standards Organisation ISO recognises X 25 1980 as a suitable basis for implementation of the lower 3 layers of the OSI model X 25 1984 88 is fully OSI compatible A 2 2 How the X 25 Protocol Works X 25 defines the protocol between Data Terminal Equipment DTE and Data Circuit terminating Equipment DCE which are operating in packet mode The DTE may be a terminal or a host computer The DCE may be the entry point into a PSE or into a network such as PSS X 25 describes three functional layers physical link and packet and is consistent with the ISO OSI model A 2 2 1 Level 1 the Physical layer This describes the physical functional and electrical characteristics of the line connecting the DTE and DCE Level 1 references the X 21 and X 21bis standards X890 304751 Issue 1 A 3 Rev 0 X 21 is designed for high speed access to digital networks X 21bis is equivalent to V 24 and V 35 and makes allowance for modems on analogue networks Level 1 provides a full duplex synchronous facility at speeds described by X 1 up to 64 000 bps A 2 2 2 Level 2 the Link Layer This describes the procedures across a line between a DTE and DCE and makes use of the services provided by the Physical layer Level 2 uses the Balanced Link Access Protocol LAPB which is consistent with ISO s HDLC procedures Level 2 transfers frames of information between
6. LCNs PVC none PVC none Incoming none Incoming none Two Way 1024 1031 Two Way 1024 1031 Outgoing none Outgoing none Level 3 Window Size 2 2 Default Max Packet Size 128 128 Logical Interface X 25 1980 DTE X 25 1980 DTE X890 304751 Issue 1 3 30 Rev 0 Power on the PSE ensuring the system disks are correctly inserted Wait for the node to boot up then press RETURN on the manager terminal to wake it up If the node has not come up properly refer to Section 3 7 1 below Log on to the Node Manager by entering L RETURN your user name and password Note that the password will not be echoed Select the Configuration Node Configuration node status display Screen This will allow you to ensure that the manager can see the XI Ms to which we are trying to connect Slots 1 and 4 should contain a module of type XIM1 The XIMs should bein the Operational state Note that state here refers to the software state of the whole XIM If the XIMs are not Operational refer to Section 3 7 2 below Press PF4 to return to the main menu Now allocate the logical ports Select the Configuration Logical Port Allocation logical physical port display Screen As no logical ports have been set up yet the message No logical port numbers have been allocated will be displayed and the screen will not be entered Goto the create a new logical port screen and type in the logical port number for PAD 1 Use logical port
7. e Timeout Period T1 This is the Information Frame Timeout value The Level 2 code uses this timeout to detect the loss of transmitted frames or their acknowledgments Valid values are between 1 and 200 tenths of a second with a default of 36 Note that the T2 timeout period Response Timeout value is automatically set to two thirds of T1 with a maximum of 1 second e Out of Service Timeout T3 If the port has been in line idle channel state for timeout period T3 then the L2 DCE will notify the higher layers that the link is down Valid values are between 0 disabled and 600 tenths of a second with a default of 200 e Idle Link Timeout Period T4 If no frames are received for timeout period T4 then the Level 2 code will send an RR frame to ensure that the link is operational The PSE ensures that T4 is set to a value greater than T1 and less than T3 Valid values are between 0 disabled and 250 tenths of a second with a default of 100 e Maximum Frame Retry Count N 2 This value is the number of times the Level 2 code will try polling the connected device following a T1 timeout before it gives up and sends an SABM frame totry to put the link back into a known state Values are from 1 to 20 with a default of 10 X890 304751 Issue 1 3 14 Rev 0 Extended sequence numbering If this parameter is set to YES then the Level 2 code will support extended frame sequence numbering modulo 128 over the link Otherwise basic frame se
8. C 9nn7 nnnn allow d nnnn nnnn bar This will produce the following results a Any calls from network 1234 destined for network 5678 will be allowed b Any calls from network 1234 destined for networks whose DNIC starts with a 5 e g 5010 will be barred c Any calls from networks whose first digit is 9 and last is 7 e g 9307 will be allowed d Any other calls will be barred If this entry is missing then these calls will be allowed 3 9 6 X 25 Gateway To configure a port as an X 25 gateway the following options must be selected as specified These are in addition to the port configuration as defined in Section 3 4 The following options must be selected on the configuration Port configuration X 25 X 75 port configuration Data Link level Screen X 75 support No Protocol Option Normally 1 Must be 2 if connecting to British Telecom s PSS The following options must be selected on the configuration Port configuration X 25 X 75 port configuration Network level Screen Profile dentifier Relevant X 25 profile or relevant network profile Tymnet Telenet Uninet or PSS if connecting to British Telecom s PSS Do not select the CCITT X 75 profiles X890 304751 Issue 1 3 47 Rev 0 The following options must be selected on the configuration Port configuration X 25 X 75 port configuration User facilities screen Gateway Yes DNIC of Local RPOA DNIC of attached network PDN Gateway Table entries or Address A
9. LCN The packet header identifies the type of packet As described below there are special and different types of packet used for establishing and removing a virtual circuit There are other types of packet used for exchanging data and control information over an established virtual circuit Different types of packet hold different types of user information Packets used for establishing a virtual circuit hold user information such as the X 121 addresses of the calling and called DTEs requests for User Facilities and transparent User Data Packets used for resetting or clearing a virtual circuit hold useful information about why the circuit was cleared e g Blind Buffer Queue Overflow in State C2 Packets exchanged over an established virtual circuit hold User Data or flow control information X890 304751 Issue 1 A 6 Rev 0 A 2 3 Procedure for a Switched Virtual Circuit Figure A 4 shows an example of the normal procedures for establishing using and clearing an SVC DTE DCE DCE DTE REMOTE Call request 3 Incoming call 4 Call accept Call connected Data Data Data Data Clear request Clear indication Clear i confirmation Clear confirmation Figure A 4 Call Procedure Using an SVC 1 A DTE requests an SVC by sending a Call Request packet to the DCE The DTE allocates the highest available logical channel number to the call The Call Request packet holds the X 121 address of the called DTE It also holds requests fo
10. One way logical channels outgoing Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes One way logical channels incoming Non standard default packet sizes 16 32 64 128 256 512 1024 2048 4096 Non standard default packet window sizes Default throughput class assignment Flow control parameter negotiation Throughput class negotiation 1 16 Closed User Group CUG 1 17 1 18 CUG outgoing incoming access 1 19 1 20 Incoming Outgoing calls barred within a CUG Y Y Y Y Y Y Y Y Y Y Y Y Y Y Yes es es es es es es es es es es es es es es es Y Bilateral CUG Not Supported Not Supported Bilateral CUG with outgoing access Not supported Not supported Table A 1 X 2 Subscription User Facilities X890 304751 Issue 1 A 9 Rev 0 Supported by Xpress r FastSelectAccepiance ve ve 124 Reverse Charging Acceptance ve Ye 4 25 Local charging prevention Ye ve 1 26 nsw Yes e i Yes i i es Call redirection Call deflection subscription 1 33 TOA NPI Address Subscription Not Supported Not Supported 1 34 Direct Call Not Supported Not Supported Table A 1 continued X 2 Subscription User Facilities Yes Yes Yes Yes N Yes Yes Yes Yes Yes 1 29 RPOA Subscription 1 30 Hunt Group X890 304751 Issue 1 A 10 Rev 0 X 2 Index Supported by Xpress 7 number 2 1 Flow control parameter negotiation 2 2 Throughput class negotiation
11. Y Y CUG selection Y Y Y X 2 User Facility 1984 88 CUG with outgoing access 2 5 Bilateral CUG selection Not Supported Not Supported 2 6 Reverse Charging es Yes es s es Yes NUI Selection Yes Yes 2 9 Charging Information Yes Yes 2 10 RPOA selection No Yes Call Deflection Selection Yes Yes 2 12 Call redirection or call deflection No Yes notification 2 13 Called line address modified No Yes notification Transit delay selection amp indication Yes Abbreviated address calling Yes Table A 2 X 2 1988 Per call User Facilities X890 304751 Issue 1 A 11 Rev 0 A 2 5 Additional Notes about Xpress Support of Some X 25 1 Facilities Non standard default packet sizes The following sizes are supported 16 32 64 256 512 1024 2048 and 4096 octets The standard default size is 128 octets Xpress does not constrain the packets to be the same sizes for each direction of data transmission or at each end of a VC Note that use of the larger packet sizes may lead to a shortage of packet buffers on UPMs with only 1 Mbyte of memory Non standard default window sizes window sizes between 1 and 7 are Supported or between 1 and 127 if extended sequence numbers are selected The standard default size is 2 Xpress does not constrain the window sizes to be the same for each direction of data transmission or at each end of a VC Default throughput class although a port can be configured with a default
12. all Xpress Virtual DTEs see Chapter 6 b Digits 1 to 4in the tagged regular expression will map onto this node and ports 11 to 14 eg 1234 112 000 3 will map onto this node port 13 c Digits 4 to 9 in the tagged regular expression will map onto this node and ports 24 to 29 eg 1234 112 002 7 will map onto this node port 27 d Any address whose first 6 digits match with 1234 56 will be forwarded to node 13 e g 1234 56 789012345 and 1234 56 will be forwarded to node 13 e Any address whose first 6 digits match with 1200 23 will be routed to hunt group 8001 on the current node e g 1200 23 789012345 and 1200 23 will be routed to hunt group 8001 on the current node X890 304751 Issue 1 4 19 Rev 0 f Any three digits in the first tagged expression will map onto the node number and any four digits in the second tagged expression will map onto the port number e g 9876 379 0923 will be routed to node 379 If this example Address Analysis Table is at node 379 then the call will be routed to port 0923 g Any address whose first four digits match with 7639 will be routed to the PDN Gateway 2342 e g 7639 1234 5678 will be routed to PDN Gateway 2342 h Any Xpress network address destined for nodes 100 to 199 will be routed as if they were destined for node 100 This enables nodes to be grouped e g nodes 100 to 199 are in London with only one routing table entry for node 100 required i Ma
13. m 2 User ID length NUI Selection m 3 facility 4 16 byte Calling User ID n nE Call User Data n 4 4 byte User Token Figure E 2 Reselection PAD Message F ormat 3 o N Oo Oo ABA WO DPD The coding of the Reselection PAD message is defined in CCITT Recommendation X 29 1984 The reselected DTE address length field has value zero and the re selected address is absent in the message format given above The coding of the Call Deflection Selection Facility fields is as defined in CCITT Recommendation X 25 draft 31 March 1988 The coding of the NUI Selection Facility fields is as defined in CCITT Recommendation X 25 1984 X890 304751 Issue 1 E 5 Rev 0 E 4 Xpress Port Configuration This section summarises the PSE port configurations necessary to support a hypothetical ACS system See Figure E 1 which shows the topology of our hypothetical network Notes on the network 1 There are both protected and unprotected user X25 ports 2 There are several services available to the users of both protected and unprotected ports 3 There is one Master ACS and two Slave ACSs The slaves support two user communities we shall call A and B The users in community A can only access the three services host 1 to host 3 under ACS control they have no other access at all The users in community B have access to hosts 1 and 2 under ACS control They may access host 3 directly In addition community B users have free access to PS
14. node can support up to 100 CUGs A DTE can subscribe to 100 CUGs See Section 4 6 for a full description Charging Information the charging information provided is in terms of call duration and number of segments transferred Hunt Group this is an X 25 1988 User Facility See Section 3 9 for a full description D bits a port is configured for one of three classes of D bit support NoD bit use allowed D bit use on request on a per call basis D bit used on all calls D bit modification Abbreviated address calling Xpress provides this facility by means of address translation and or Flexible Addressing See Section 4 4 Called line address modified notification this is an X 25 1988 User Facility which Xpress provides for the support of Call Redirection Deflection CUG with outgoing access this is an X 25 1988 User Facility Transit Delay selection and indication Xpress does not take into account the selected transit delay when it routes calls nor does it measure the actual delay Xpress sets the Delay Indication to the 1988 code unknown Call Deflection Selection This facility is supported as per X 25 1988 This support is described in detail in Appendix E Any connected device may use this facility if it is capable of doing so Call Deflection Data Transfer Selection This non standard extension of Call Deflection Selection allows the ACS to deflect a call which has reached a data transfer state This
15. so in most instances the parameters found on this menu can be left at their default values However one of the parameters Priority Class Profile may need to be configured if Congestion Monitoring and Control is in use on any trunk ports elsewhere in the network For further details refer to Section M 2 2 A full description and configuration guidance for Congestion M onitoring and Control can be found in Appendix M 3 4 7 Error Monitoring Error Monitoring and Control can improve the efficiency of the Xpress network as it enables a node s routing process to avoid using links or trunks with high error rates When initially getting a node working it is not necessary to configure this feature Later on its use should be considered if it is known that the lines carrying the network s links and trunks are generally of a poor quality This feature works by monitoring the mean error rate on a port and temporarily closing it whenever an unacceptably high rate of errors is detected on it The port behaves just as if it had gone down all existing calls using it are cleared and the equipment at the remote end sees the X890 304751 Issue 1 3 22 Rev 0 HNK aS DeIng Gown calis attempting to teave tne node via tne ciosed port will choose an alternative port providing one exists this would be the case for example if the port belongs to a hunt group It can be arranged that a port previously closed in this way will automatically be reinstate
16. 1 or be in the 14 character Xpress Addressing format characters 1 4 must be 1100 characters 5 7 must match the node number this is displayed on the top line of the manager terminal e g 001 characters 8 11 must match the logical port number of the called PAD characters 12 14 must be understood by the PAD as a port selector e Bad calling address Either the PAD must supply a calling address in the same format as the called address e g 11000010001000 for PAD 1 and 11000010002000 for PAD 2 or an address translation must be set up on the calling PAD s port See Section 4 5 2 e Bad facilities Make sure the facility settings on the Configuration Port Configuration X 25 Port Configuration user facilities Screen match the requirements of each of the PADs Particular cases to watch are flow control parameter negotiation and basic extended format selection The other parameters are much less likely to cause trouble You will have to look up the requirements of the PADs in their documentation e PAD async port problems Check that there is a terminal connected to the async port which you are trying to call on the PAD Make sure the PAD can see it e g type STAT RETURN if it is X 28 compatible and see if it responds FREE or similar e LCN Range Error This is the most likely cause of a call disappearing If the PSE port s LCN range allocation does not match that which the PAD is expecting the XIM or the PAD depending on who made the
17. 1 6 Error Monitoring Parameters These parameters should be configured if Error Monitoring and Control is to be used on a trunk port They are identical to the parameters for X 25 X 75 ports refer to Section 3 4 7 for detailed information Enabling this feature allows the system to close a port temporarily whenever an unacceptably high rate of line errors is detected on it The port is treated just as if it had gone down with all the affected calls being internally cleared and re established via an alternative next hop trunk port if possible Error Monitoring and Control should be considered for use on a trunk port if itis liable to have an unacceptably high error rate such as 10 or higher due to the use of poor quality lines and thereis an alternative next hop trunk port secondary or tertiary configured X890 304751 Issue 1 4 7 Rev 0 The most recent error rate measurement can be readily inspected on the Configuration Node Configuration Detailed Link Status Display Screen e Auto Reroute Interval This is described in detail in Section 4 3 1 4 and is found on the Configuration Port Configuration Trunk Port Configuration Network Level menu This should be configured for use on any trunk ports that are secondary or tertiary next hop trunk ports if Error Monitoring and Control is in operation on the primary port The Auto Rerouting process regularly clears all calls that are using the port as their secondary or terti
18. 1 A 16 Rev 0 Level 2 the Link Layer Xpress supports Level 2 of X 75 interfaces in the same way as for X 25 interfaces see A 4 2 except for an X 75 1980 interface configured for extended sequence numbering For extended sequence numbering the X 75 1980 protocol requires a two byte control field for unnumbered frames whereas X 75 1984 and X 25 1984 88 specify a single byte control field Xpress provides a configuration option which the operator may use to specify the required variant of Level 2 Level 3 the Packet Layer The differences between X 75 and X 25 at the Packet Layer are mainly either extensions e g extra fields in X 75 call set up packets or less rigorous error checking than X 25 e g discarding unexpected packets rather than issuing clears resets Network Utilities The main difference between X 75 and X 25 is that X 75 call set up and clear down packets hold an additional field the Network Utilities field The Network Utilities are located between the Address field and the User Facilities field and have a maximum size of 63 bytes The Network Utilities field holds user facilities which are relevant to X 75 interfaces and consist of X 75 Network Utilities and X 25 User Utilities as described below X 75 Network Utilities a Transit Network Identification Codes TNICs These record the identity DN1IC of every transit network and are held in call set up clear down packets Note that the DNICs of th
19. 1 to 3 on frame relay virtual physical link 0 7 1 T0071 multiplexed over physical Node 3 link 0 7 0 to FPX and identified by local DLCI 17 Physical Connections Sekeni Logical connections Figure 4 1 Example Frame Relay Trunk Configuration Figure 4 1 shows how the trunks are physically connected The trunk to node 2 is on virtual physical port 0 7 0 logical port T0070 and that to node 3 is on virtual physical port 0 7 0 logical port T0071 The various configurable parameters are as follows Frame Relay Physical Port Configuration Physical Level All parameters compatible with defaults other than clocking which is provided by FPX2000 at 256 kbps and physical interface variant which is X 21 Core Data Link Level All parameters compatible with defaults other than R bit support which should be enabled X890 304751 Issue 1 4 10 Rev 0 Trunk Port Configurations Data Link Layer All parameters compatible with defaults other than DLCI which is 16 for port T0070 and 17 for port T0071 Node 1 is acting as DCE in both cases Network Layer All parameters compatible with defaults other than logical circuit number range which is 1 to 128 for both trunks Node 1is acting as DCE Step 1 Select the correct Frame Relay application for slot 7 1 Log onto the manager N Select Configuration Module configuration Edit module parameters Type 7 RETURN to select slot 7 A Ww Type a RETURN to e
20. 6 6 4 4 amp 4 5 4 4 3 4 5 3 Rev 0 Section no O utgoing address translation S ource calling address 4 5 4 translation D estination called address translation P DN gateway specification 3 9 CrJeate PDN gateway E dit PDN gateway D elete PDN gateway L ist PDN gateways A ddress analysis table configuration 4 5 1 D NIC Barring Table for X 75 3 9 5 S statistics 5 10 Modify report 5 10 5 L ink statistics section 5 10 6 Mjodule statistics section 5 10 7 I ntra node communication incs statistics section 5 10 8 C ontrol report 5 10 9 D isplay port statistics 5 10 1 P hysical level 5 10 2 Frame relay C ore level 5 10 3 Frame relay L MI level 5 10 4 F rame level 5 10 5 Pa cket level 5 10 6 U tilities A ccess utilities 5 1 C hange password 5 1 1 U ser access specification 5 1 3 C reate user 5 1 3 D elete user 5 1 3 E dit user attributes 5 1 3 L ist all users on this 5 1 3 node T ype specification 51 2 A larms and warnings St atistics monitoring U user access specification Sy stem utilities P hysical configuration B illing specification R outing specification C lock utilities Date change T ime change X890 304751 Issue 1 l 3 Rev 0 Section no Di sk utilities 5 3 m D isk copy 5 3 2 Fo rmat disk 5 3 1 L ist contents of disk 5 3 3 Verify disk 5 3 7 Fi le copy 5 3 4 R emove f
21. 75 Application Port Configuration User Facility Local NUI selection where the Validate option should be selected X890 304751 Issue 1 P 2 Rev 0
22. C Xlink Trunk These 4 PADs can support gt Broadcast data flow up to 32 clients 4 ABS Server on slot 4 Figure G 5 Example of Multi service ABS X890 304751 Issue 1 G 15 Rev 0 G 5 Capacity and Performance G 5 1 Sizing The overall capacity of the ABS is limited by the maximum number of clients that can be connected to a single server and the maximum number of servers in a network Every X 25 card in a node carries an ABS server as does the manager Each X 25 card s server can handle up to 32 clients and the node manager s server up to 64 it is important to note that in each case this call limit is shared with all other virtual DTE calls For example if there is an existing call to the load generator on a given X 25 card then only 31 ABS clients can be served by that card while the load generator call exists This is especially relevant to the manager card which uses the virtual DTE mechanism for access to the Mini PAD remote printing and other systems Note that cards running the Xpress Kernel plus an application do not have an ABS server It is therefore recommended that the ABS server on the node manager be used only to broadcast to other servers within the same node and not to support clients directly This will then leave sufficient capacity in the manager s virtual DTE system for node management functions If configured as suggested above the workable maximum number of clients per node is th
23. Comment 1 wizard wand Can write to any screen 2 engineer case For use by a Cray engineer The engineer cannot change the Access Utilities screens and the Billing Configuration 3 super nms Local Supervisor 4 op op Local Operator Can write tothe System Utility Alarms and Warnings screens X890 304751 Issue 1 5 4 Rev 0 5 2 Clock Utilities This menu allows you to change either the current date or the current time The current time and date is displayed in the top right of the manager screen and on reports output on the printer The current time and date are maintained by a real time clock chip on the Utility module This unit will maintain the correct time and date during system downtime for up to a maximum of about one week If the system does lose the time and date on restart a default time is assumed midnight 1st J an 1900 After a date time change it may take a few seconds for the MMI display to be updated 5 2 1 Change Date To change the date enter day month year in the format DD MMM YY eg 12 Nov 90 5 2 2 Change Time This screen enables you to change the time that is displayed in 24 hour clock form at the top of each screen To change the time enter hours and minutes in the format HH MM X890 304751 Issue 1 5 5 Rev 0 5 3 Disk Utilities This menu lets you select one of the disk or file utility functions The disk identity is synonymous with the drive identity i e disk a refers to drive a and
24. Control IS operating In a network INe associated resets and potential loss of users data can be prevented by enabling Network Data Integrity at the user ports M 2 3 Configuring Trunk Ports on Secondary Routes e Auto Reroute Interval This is described in detail in Section 4 3 1 4 and is found on the Configuration Port Configuration Trunk Port Configuration Network Level menu If Congestion Monitoring is in operation on a primary next hop trunk port then auto rerouting should be considered for use on each of the trunk ports that comprise the secondary or tertiary route The Auto Rerouting process regularly clears all calls that are using the port as their secondary or tertiary choice This forces them to re establish with the result that these previously displaced calls can be periodically returned to their primary route if it is available M 2 4 Trunks to Pre Version 9 Nodes The system needs to assign a priority to a call request at the first Version 9 node it encounters Although this will normally occur at the port of entry to the Xpress network i e an X 25 X 75 port as described earlier it may need to be at a trunk port if the trunk attaches to a node running an older version of software The Priority Class Profile parameter Configuration Port Configuration Trunk Port Configuration Congestion Monitoring Should be configured on this trunk port so that it correctly maps the call s Throughput Class Facility to the appropriate
25. Login command the Node Manager will prompt you for the entry number of the required application The Node Manager will display a message Attempting to connect to application to indicate that it is setting up a call to the selected application If the login attempt is successful then the screen is cleared and dialogue with the application begins If you have configured a suitable username and or password then the application may not need to prompt you for a username and or password Otherwise the application will either reject the attempt immediately or prompt you for a username and or password The Node Manager will log a WARNING event whenever an attempt to call an application fails Once you are logged onto the application the PSE Node Manager will suppress its usual display of status information on the top and bottom lines of the screen see Section 2 2 4 so as to provide the application with access to the whole screen However the Node Manager will maintain the usual inactivity timer It will appear busy to any login attempts 2 3 3 Logging Out from an Application When you log out from an application manager the PSE Node Manager will present you with the Application Management screen The Node Manager will also resume its usual display of the top and bottom lines of the screen see Section 2 2 4 You can now login to another application or select another Node Manager screen If your connection to the PSE Node Manager is clear
26. Of Caras along with a Gate application then tnere iS not enough disk space available to do a download in one single operation If anode has an application installed then the download operation will result in the need to re install the application Hence the application distribution disk should be at hand 1 Login to node 30 and remove any unwanted files on drive b e g dump files to make as much space available as possible 2 Logout of node 30 Login to node 1 From the utilities Disk Utilities File Copy screen copy the boot data and a limited number of load files from the local drive a to the remote drive b of node 30 The from file will be a a z atL and the to file will be b30 The pattern matching from filename will match the data v8boot and lower case load files Once the operation has completed logout of node 1 6 Login to node 30 and list the contents of drive b to check that all the files are present 7 Now copy the newly downloaded files on driveb todrivea The from file will be b a z atL and the to file will bea 8 Now tidy up node 30 driveb removing the files just downloaded Use the remove command with the following file name Remove b a z atL 9 Now you have to copy the node manager load file However the node manager is stored on drive b and this is the temporary download disk Hence the filenames clash and you would overwrite the current node manager One way to get around this is t
27. Primary Gateway Node 2 port 0002 Secondary Gateway Node 1 port 0002 Tertiary Gateway Node 3 This node supports community B who are served by ACS slave 2 Primary Gateway Node 3 port 0003 Secondary Gateway Nodel port 0002 Tertiary Gateway 9 This arrangement will cause user calls to be routed to the requisite slave ACS in the first instance and then to the master ACS if the slave is inaccessible E 4 2 User Port Configuration All protected user ports Node 2 ports 0003 and 0004 Node 3 ports 0001 and 0004 must have the Call Deflection Referral facility set to YES Now we must set up the user ports Incoming Called Address Translation ICAT tables to route user calls to the correct ACS Node 2 port 0003 This port is used by members of User community A who have no free access in the network and are handled by ACS Slave 1 Thus we set up the ICAT on this port to route all incoming user calls to Slave 1 X890 304751 Issue 1 E 7 Rev 0 ICAT Match Address ICAT Substitute Address nnnn nnn nnnan nnn 9998 000 0000 1 NULL 9998 000 0000 1 This maps any called address to the DNIC of ACS Slave 1 and thus any user call will go to Slave 1 or to the Master Slave 2 if Slave 1 is unavailable Node 2 port 0004 This port is identical to port 0003 apart from the fact that the users are DCX network users They arestill in community A Hence we configure the ICAT table for this port the same as for
28. Remote Node Management in Version 2 to 5 of the software WNrF X890 304751 Issue 1 6 2 Rev 0 4 Billing Call Data 5 Centralised Printing 6 Mini Pad From Version 6 onwards Ts 8 9 Not used Diagnostic Virtual DTEs see Section 6 1 2 X890 304751 Issue 1 6 3 Rev 0 6 2 Module Crashes If a module develops a fault which stops it from operating it is said to have crashed A crash may occur because of a hardware or software problem and the action the system takes is different for each If a module crashes because of a hardware fault it will not attempt to dump nor will it automatically reload It will move into Call Operator state and remain there until the operator triggers a reload Thisis because a hardware fault is probably not transitory and reloading the board s software will therefore not help If a module crashes because of a software fault it will normally attempt to dump any useful software fault diagnosis information to a file on disk see Section 6 3 It will then automatically reload the software and move into its operational state again Dump followed by auto reload is the normal default sequence However either dumping or auto reloading can be disabled on a per module basis using the Configuration Module Configuration Edit Module Parameters screen If the operator does disable auto reload then on crashing the module will move to Call Operator state and remain there until the operator triggers a
29. Six protocol ports via physical ports 0 to 5 4 X 25 2 FR Four protocol ports via physical ports 0 to 3 plus four FR virtual physical ports 4 and 5 multiplexed over FR physical port 4 3 X 25 2 FR Three protocol ports via physical ports 0 to 2 plus three FR virtual physical ports 3 to 5 multiplexed over FR physical port 3 2 X 25 4 FR Two protocol ports via physical ports 0 and 1 plus four FR virtual physical ports 2 to 5 multiplexed over FR physical port 2 6 FR O X 25 Six FR virtual physical ports 0 to 5 multiplexed over FR physical port 0 4 FR 2 X 25 Four FR virtual physical ports 0 to 3 multiplexed over FR physical port 0 plus two protocol ports via physical ports 4 and 5 3 FR 3 X 25 Three FR virtual physical ports 0 to 2 multiplexed over FR physical port 0 plus three protocol ports via physical ports 3 to 5 2 FR 4 X 25 Two FR virtual physical ports 0 and 1 multiplexed over FR physical port 0 plus four protocol ports via physical ports 2 to 5 Table 3 1 8325 Example Native Applications X890 304751 Issue 1 3 7 Rev 0 APPLICATION CARD S DESCRIPTION Node Manager UPM3 UM The node manager software Crash Dump UPM3 UM Dumps the node manager should it fail Do NOT select this application manually X 25 K 75 UPMx XIM Four protocol ports via physical ports 0 to 3 3 X 25 1 FR UPM3 XIM_ Three protocol ports via physical ports 0 to 2 plus one FR virtual physical port 0 multiplexed over FR physical port 3 2 X 25 2 FR UPM3
30. a config the to file will be b1 6 Logout of node 30 7 Login to node 1 and list the contents of drive b to check that all the files are present 8 Remove the disk in drive b and write protect it This process can be repeated for all the nodes in the network if desired It is not meant as a replacement for the NMC Upload Download system just an alternative procedure which you may choose L 3 2 Remote Software Version Download 8325 Before you attempt any of the following operations you must have backups available of all the disks on all of the nodes that you will be accessing This procedure can be used to update the software of an 8325 node remotely Assume the following Node number 1 is the central node and is running Version 8 2 A remote node number 30 is running Version 8 1 and needs to be upgraded to Version 8 2 All nodes are 8325s The operator is working from the Mini Pad at node 1 1 Login to node 30 and remove any unwanted files on drive b e g dump files to make as much space available as possible X890 304751 Issue 1 L 4 Rev 0 2 Logout of node 30 Login to node 1 From the utilities Disk Utilities File Copy screen copy all the software files from the local drive a to the remote drive b of node 30 The from file will be a aLt and the to file will be b30 The pattern matching from filename will match the data L load files and the v8xboot files 5 Wait for the operation to complete It will
31. a management process within the Node Manager UM 9999 This addresses the virtual DTEs on the XIM through which the call entered the node X890 304751 Issue 1 6 1 Rev 0 Each virtual DTE has a specific sub address which identifies it The sub addresses are 0 The packet sink Simply accepts all data that is sent to it 1 The packet echo Echoes back all data that is sent to it 2 Thepacket source An unlimited supply of data Will sink all data that is sent to it and continue to send you a message 3 Memory dump For security reasons this is only available on development versions of the software 4 Software Trace 5 Asynchronous Broadcast Service Server access see Appendix G 6 1 3 When It Should Be Used Ideally each time a new piece of equipment is connected to the network calls should be placed to the virtual DTEs to ensure that no basic configuration problems have occurred Calls should also be placed to destinations across the network in order to determine any problems with address translation that may have been introduced by adding new equipment 6 1 4 Node Manager Virtual DTEs To access the Node Manager Virtual DTES make an X 25 call to the following addresses 1100 nnn 900 ppp where nnn is the node number ppp is the Virtual DTE subaddress This is only used by the diagnostic Virtual DTEs is the Virtual DTE link number Node M anager NMC interface Not used These were used by
32. added to provide a secure way to update files when transferring new versions between nodes It can also be used in general to rename files for backup purposes for instance L 1 5 Node Restart All the cards in a node can be made to restart simultaneously by using the new node restart command This command is present on the Configuration Node Configuration Edit Node Configuration screen By choosing the Node State command and entering restart followed by PF1 submit the node will be restarted Use this command with care X890 304751 Issue 1 L 2 Rev 0 L 2 Security Considerations The ability to copy files between nodes opens up the ability to accidentally corrupt a node remotely In addition if not used correctly a security loophole can be created The following sections discuss the areas of security that are affected by the introduction of remote software download L 2 1 File Corruption All file copy operations are verified by using a 32 bit CRC to detect errors after the file has been transferred If an error is detected the file copy operation aborts with a warning that a CRC error was detected The file however is still left on the remote node but obviously it is corrupt If you were updating a load file obviously this could be fatal You should never attempt to update a node s load files by over writing them You should always use one of the following methods Transfer the new file with a temporary new filename and t
33. always be set to FFFF since Xpress cannot accurately calculate this value 5 Packet sizeis 4 6 DAWPOANAUBW AH OWPVOANDU 16 octets 32 octets 65 octets 128 octets 256 octets 512 octets 1024 octets 2048 octets 4096 octets hroughput class is 75 bps 150 bps 300 bps 600 bps 1200 bps 2400 bps 4800 bps 9600 bps 19200 bps 48000 bps 64000 bps X890 304751 Issue 1 C 5 Rev 0 X890 304751 Issue 1 C 6 Rev 0 Appendix D Closed User Group Call Permissions D 1 Example Network Figure D 1 shows an example network to which five DTEs are connected Two CUGs are defined Table D 1 gives the details of the CUG membership of the DTEs Figure D 1 Example of CUG Groupings DTE CUG Subscription and Membership A Incoming access CUG 1 with incoming calls barred B No external access CUG1 C Incoming access CUG 2 CUG 1 with outgoing calls barred D Outgoing access CUG 2 E No CUG subscription Table D 1 CUG Membership X890 304751 Issue 1 D 1 Rev 0 These settings produce the call permissions shown in Table D 2 DTE Can Make Calls to Can Receive Calls from A B C D E B C A C D A B D E E A C D X890 304751 Issue 1 Table D 2 CUG Call Permissions D 2 Rev 0 D 2 Call Permissions and Prohibitions A detailed list of permissions and prohibitions for the example network is given in the following sections for DTE B in CUG 1 DTE D in CUG 2 and DTE E which is not in any CUG D 2 1 CUG 1 Permiss
34. and time The bottom line of the screen always shows the current counts of PSE alarms and warnings more about these in Chapter 5 Between these two constant displays are listed the commands for the menu you are using To select a command simply type in the command name followed by a RETURN This gets you to the next menu in the sequence The first letter s of each command is highlighted and you can select a command by typing in just the highlighted portion This is the minimum that can be entered for that command to be identified uniquely 2 2 5 Use of the Function Keys The VT 100 programmable function keys are used by the Node Manager to control the menus PF3 and PF4 may be used at any menu and have the following meaning PF3 At any menu pressing the PF3 function key returns you tothe previous menu unless of course you are at the main menu PF4 At any menu pressing the PF4 function key returns you to the main menu The PF1 and PF2 function keys are also used at certain menus 2 2 6 Online Help At any time pressing causes help text to be displayed The window displays text explaining what the menu does and how to use it X890 304751 Issue 1 2 6 Rev 0 2 2 7 Logging Out The main menu is the only one from which you can logout Entering the command L logs you out from the Node Manager and clears the X 25 call The Node Manager also logs out automatically if the terminal is idle for about ten minutes or
35. byte packet per second to the maximum of 32 clients while leaving a reasonable amount of processing power approximately 50 for a standard power XIM and approximately 80 per high power XIM available for non ABS calls X890 304751 Issue 1 G 17 Rev 0 G 6 Diagnostics and Error Handling Two events have been defined to help set up tune and fault find the ABS Host Call Failure Alarm and Data Being Discarded Warning The first is generated if a server is unable to make the connection to the configured host address or if the host call once established is cleared by either the network or the host itself The format of the event is Broadcast server host call cleared Cause CCC Diagnostic DDD Bay 0 Slot SS where CCC DDD represent the X 25 clear cause and diagnostic codes in decimal and SS represents the slot number of the server whose host call was Cleared It is worth noting that a server without the requisite AAT entry for its host call s called address will be cleared with a cause and diagnostic code combination of 128 67 Xpress Network Clear Invalid Called Address The second event is generated the first time data is discarded for any client call and then subsequently every time 1000 data packets have been discarded for that call The format is Broadcast server discarding data for the client at X 121 address XXXXXXXXX Packet s discarded NNNN Bay 0 Slot SS where XXXXXXXXX is the X 121 address up to 15 di
36. can be many servers in a network operating either independently or co operatively The way the system is configured determines the server to which the client connects to pick up the required broadcast data The host and clients must be asynchronous DTEs directly or indirectly connected to the Xpress network via a suitable PAD Examples of suitable hosts and clients are Asynchronous terminals connected via a triple X PAD i e Cray 8160 or DCX X PAD Applications within PCs workstations connected via an X 25 card with suitable PAD software LAN workstations connected via a suitable X 25 triple X gateway Generally synchronous access devices use a transport layer protocol requiring two way communication between end systems This means that X890 304751 Issue 1 G 1 Rev 0 tney are unsuitable tor oroaacast applications aS tne poroaacast aata TIOW IS necessarily unidirectional from host to clients X890 304751 Issue 1 G 2 Rev 0 G 2 Using a Single ABS Server G 2 1 Client Access to the Server A client wishing to receive broadcast data must input the general address of the broadcast server The general address in a given card is the address of Virtual DTE number 5 on that card The general address format is 1100 nnn 9ss9 005 where nnn the node number and ss the slot number of the card containing the required ABS server For example a terminal connected toa PAD wishing to connect tothe ABS server in node 140 slot 7 mak
37. certain ports if Congestion Monitoring and Control is being used within the network For details see Appendix M e Flow Control Parameter Negotiation If this parameter is set to YES negotiation of call packet and window sizes will be enabled This means that the software will allow both incoming and outgoing calls to use the negotiation procedures to change the values from the default if necessary Default is NO e Fast Select Call Acceptance The PSE always accepts incoming Fast Select calls but will only send them out on a port which has Fast Select Acceptance enabled Thus for a Fast Select call to succeed the target device s port must have this parameter set toYES Default is NO X890 304751 Issue 1 3 19 Rev 0 e Extended Format Selection X 25 only When set to NO this parameter will cause any Call Request with facilities to be cleared Default is YES e Gateway port This parameter must be set to YES when the port is being used as a gateway between the PSE and another public or private network e g PSS Thedefault is NO e D bits D bit Delivery Confirmation bit usage is explained in Appendix A The available values are Enforced On Request and Disallowed These indicate that D bits must be used may be used and must not be used respectively Enforced is equivalent to the X 25 D bit modification facility Default is On Request Second Page e Call Deflection Allowed X 25 only Call Deflection isa CCITT 1988 facility whic
38. data A packet consists of user data contained within an envelope of control and address information Packet Switching provides on demand multiplexing of multiple connections over a single circuit thus allowing optimum utilisation of that physical circuit Because the Packet Switching connections are not based on dedicated physical circuits they are called virtual circuits Packet Switching supports two types of virtual circuits e Switched Virtual Circuits SVCs are set up as and when they are requested and removed when no longer required e Permanent Virtual Circuits PVCs are allocated for a period of time they are always ready for use and are analogous to leased lines Packet Switching Networks impose limits on the size of packets and share out usage over all the virtual circuits Because a Packet Switching Exchange PSE switches packets only Packet Assembler Disassemblers PADs are used to allow non packet devices such as host computers and asynchronous terminals to connect to PSEs Currently two main interfaces to Packet Switch networks are being used These are X 25 and frame relay The following sections describe how the PSEs use these two systems X890 304751 Issue 1 A 1 Rev 0 A 2 The X 25 Recommendation The X 25 Recommendation defines in detail how Packet Switching Public Data Networks PSPDNs are accessed X 25 specifies the protocol across the interface between a network and equipment connected to it The
39. data bytes integer 116 long total called data packets integer 120 long total calling data packets integer 124 long total called data segments integer 128 long total calling data segments integer 132 long total called data interrupts integer 136 long total calling data interrupts integer 140 word segment size in bytes integer Table C 1 continued X 25 and X 75 Billing Information Record X890 304751 Issue 1 C 3 Rev 0 Byte Offset Size Contents 142 143 144 146 147 148 186 byte bits 7 6 Network type 0 X 25 Network not valid 1 Originating Network 2 Transit Network 3 Destination Network bits 5 4 Port Type 0 X 25 port 1 Trunk Port not valid 2 X 75 port 3 byte unused padding byte word CNIC Ox7fff indicates no CNIC present byte Number of TNICs present byte unused padding byte word 19 TNIC list long X 75 Call identifier Coding integer integer integer integer integer integer Table C 2 X 75 only Part of the Billing Information Record Notes 1 SVC count is an indicator of the number of SVC calls that have been made across this interface 2 CUG or CUG OA of ZEEE means No CUG in use 3 Protocol identifier is FE WOON OAUBWNEF O e O X 25 1980 X 25 1984 TYMNET TELENET UNINET PSS Not valid Not valid User defined profile 1 User defined profile 2 User defined profile 3 X 25 1988 X 75 1980 X 75 1984 X890 304751 Issue 1 C 4 Rev 0 4 Transit delay should
40. following conventions are used in this Appendix BitO Theleast significant bit Byte 8 bits Word 2 Bytes most significant byte first Long 4 Bytes most significant byte first Direction of information when user A calls user B calling Ay B called X890 304751 Issue 1 C 1 Rev 0 Byte Offset Size 0 Oo NOoBRW DP 17 25 33 41 42 44 46 48 49 word byte byte byte byte byte byte byte byte 8 byte 8 byte 8 byte 8 byte word word word byte byte Contents billing format version X 25 info X 25 and X 75 info year e g 88 month date hours 24 hour clock used minutes seconds bits 0 3 called address length bits 4 7 calling address length called X121 address calling X121 address called Xpress port address calling Xpress port address unused padding byte LCI at this port LCI at the remote port calling port s SVC count see Note 1 bit 7 set if SVC call unset if PVC call bit 6 set if this port is the calling port bit 5 set if the call was successful bit 4 set if the user at this port cleared the call bit 3 set if the user at this port is to be charged for the call bit 2 set if this is a management call bit 1 set if extended packet sequence numbers were used bit 0 set if the call used end to end delivery confirmation bit 7 set if reverse charging was requested bits 6 5 0 no fast select 1 fast select without restriction 2 fast select with
41. in the future users can edit these profiles In the interim P1 is a copy of the CCITT X 25 1980 Profile P2 is a copy of the CCITT X 25 1984 Profile and P3 is a copy of the PSS Profile e Default Maximum Window Size This should be set to match the Level 3 window size required by the connected device Values are 1 7 for basic sequence numbering and 1 127 for extended sequence numbering The default is 2 e Maximum negotiable window size This is the maximum window size which can be negotiated during call setup using Flow Control Parameter negotiation The values are 1 7 for basic sequence numbering and 1 127 for extended sequence numbering The default is 7 e Extended sequence numbering If this parameter is set to YES then the PSE will support extended packet sequence numbering modulo 128 over the link Otherwise basic packet sequence numbering modulo 8 will be supported The default is NO e Dial up Operation and Dial up Timeout These two parameters control the behaviour of the port dependent on whether the link to the attached device is permanently available i e leased line or permanently connected modem link or dialled i e dial up modem or ISDN TA link See Appendix K for full details of these two types of operation 3 4 5 User Facilities These parameters are used to choose which of the X 25 level 3 user facilities or X 75 network utilities will be used or accepted on this port The user facilities form is d
42. line In this state the port will not accept user calls but will accept management calls generated by Xpress nodes themselves to support management functions such as transparent login billing centralised printing etc X890 304751 Issue 1 3 25 Rev 0 c The status of configured ports can be checked at any time using the Configuration Node Configuration Detailed link status display screen This shows the descriptions port state level 2 data link status up down or errs and the number of active SVCs and PVCs for all configured logical ports d Application links are always up while their port state is on line e The level 2 link status of a frame relay virtual physical port depends largely on whether the frame relay network is successfully transferring the X 25 X 75 trunk level 2 frames between the local Xpress port and the remote device thus keeping the link protocol alive end to end There are however cases where a failure within the frame relay network is detected and the X 25 X 75 trunk links running over affected frame relay virtual physical ports go down immediately X890 304751 Issue 1 3 26 Rev 0 3 5 Virtual Circuits Packet switching connections are not based on dedicated physical circuits but on virtual circuits which in Xpress terms are logical connections between logical ports There are two types of virtual circuits e Switched Virtual Circuits SVCs which are set up as requested and removed when no longe
43. number of the program code resident in their Read Only Memories ROMs and in the U PM s Random Access Memory RAM 3 3 3 Change Module Link States Each port may bein one of three states Online Offline or Out of Service A port must be put Out of Service before you can change its configuration by using the Configuration Port Configuration X 25 or Trunk port Configuration change state of port screen The Change Module Link States option enables you to change the state of all ports on a module simultaneously 3 3 4 Module Restarts This can be used if a module has crashed and not restarted automatically or if you wish to restart the module after loading new software Before restarting a module you must stop any software running on that module then restart it from a floppy disk Modules can only be restarted when in one of the following states Operational Idle Call Operator Software Error State Restarting an Operational Module clears all calls on that module X890 304751 Issue 1 3 9 Rev 0 3 4 X 25 X 75 F rame Relay Port Configuration Once a Logical Port has been allocated you must set up the X 25 X 75 Fr configuration for that port to match the configuration of the device connected to it The port parameters are divided into seven sets Physical Level Parameters Frame Relay Core Level Parameters Data Link Level Parameters Network Level Parameters User Facilities e Congestion Monitoring e Error Monitoring This
44. on the configuration node configuration edit node configuration screen The DNIC Barring Table may need setting up to bar calls X890 304751 Issue 1 3 48 Rev 0 3 10 Frame Relay Gateways This section has been included solely to point out that Xpress frame relay interfaces are NOT gateways i e Xpress uses the services of a frame relay network to transparently carry X 25 X 75 or inter node trunk traffic between symmetrically configured devices The frame relay network does not terminate any of the protocols being transferred over it and effectively behaves like a simple direct physical interface over which multiple traffic streams are multiplexed albeit to separate destinations X890 304751 Issue 1 3 49 Rev 0 X890 304751 Issue 1 3 50 Rev 0 4 Configuring a Network 4 1 Introduction This section describes how Xpress PSEs can be linked together via trunks to form a network and how calls are routed across such a network It also describes other aspects of networking of nodes such as PVCs gateways to X 25 X 75 PDNs remote access to PSE Node Management software and automatic rerouting due to trunk failure or high error rate e Routing An Xpress network provides routing of calls across the network across any trunks and across any intermediate nodes to the specified destination node You can configure multiple paths between nodes so that the PSEs can bypass faulty or congested trunks and nodes e Address Analysis The Addr
45. or to set up a translation in the PAD to connect the call to port 3 If a client call is made to any server which does not have an appropriate AAT entry then the host call made by that server will simply fail as the raw fixed address used by the server will route the call to an illegal virtual DTE address on the same slot No other network user will be affected Once the host call is accepted any data received by the host will be duplicated and sent to each client as described above If the host sends an interrupt or reset packet these will simply be acknowledged and discarded i e they will not be broadcast to the clients If the host call initially fails or is cleared by the network or the host itself then the server will automatically generate an event to the node manager and re try the connection once per minute until it succeeds Any call user data present in the first client call is transferred into the host call along with the D bit setting of the client call The transfer of the call user data allows services to be accessed on hosts which insist on using call user data to identify the service requested Similarly transferring the D bit setting allows end to end acknowledgements to be asserted across the host call X890 304751 Issue 1 G 5 Rev 0 In the configuration in Figure G 1 the host which is an application on a mainframe is connected to port 30 on an 8325 node 1 The host expects to receive the call to pick up the
46. port 0003 Node 3 port 0001 This port is used by members of User Community B who will need a slightly more complex ICAT because they have free access to host 3 and may call PSS and WizzNet ICAT Match Address ICAT Substitute Address 3001 nnn nnnn nnn 3001 nnn nnnn nnn 2342 nnn nnnn nnn 2342 nnn nnnn nnn 1100 003 0002 nnn 1100 003 0002 nnn nnnn nnn nnnn nnn 9998 000 0000 1 NULL 9998 000 0000 1 This configuration allows any call with the DNIC of WizzNet or PSS to go straight to the requisite gateway It also allows direct calls to Host 3 Any other call is sent to ACS Slave 2 Node 1 port 0005 This port is only used for unprotected access therefore it requires no special configuration Node 2 port 0001 This is the WizzNet gateway Noincoming calls are permitted thus we set its ICAT to translate any incoming called address to NULL X890 304751 Issue 1 E 8 Rev 0 Node 3 port 0004 This is the PSS gateway We cannot be sure of the community membership of users calling via PSS This does not really matter however as we just assign ACS Slave 2 to handle these users simply because it is connected to the same node Thus we set the ICAT to ICAT Match Address ICAT Substitute Address nnnn nnn nnnn nnn 9998 000 0000 1 NULL 9998 000 00001 E 4 3 Host Ports The Host ports Node 1 ports 0003 and 0004 Node 3 port 0002 do not require any special configuration E 4 4 NMC Port The NMC port Node 1 port 1 does not require any spec
47. port which maps to an FR virtual physical port other than the one physically connected to the frame relay network then it is the physically connected port s configuration that is changed For example FR virtual physical ports 0 2 0 to 0 2 5 are being multiplexed over physical port 0 2 0 and logical port 0024 is being edited Logical port 0024 maps to FR virtual physical port 0 2 4 which in turn is multiplexed over physical port 0 2 0 and it is the configuration associated with logical port 0020 which is being edited e Maximum Frame Size N 203 This is the size in bytes of the information field of the largest frame relay frame which this port will send to or accept from the network without signalling an error This parameter is automatically cross checked with the X 25 maximum frame size i e N1 of all the FR virtual physical ports multiplexed over this port Range is from 263 to 4103 bytes with a default of 1600 bytes e Heartbeat Polling Period T391 Every T391 seconds the Xpress port will generate a frame relay Link Integrity Verification Status Enquiry message to the network which should prompt the network to return a Link Integrity Verification Status Message to confirm that the link is active Range is from 5 to 30 seconds with a default of 10 seconds e Full Status Poll Frequency N391 Every N391 Heartbeat Polls the Xpress port will replace the Link Integrity Verification Status Enquiry message sent to the netw
48. screen it lists all the applications which are installed on the PSE i e all the applications which are installed onto the system disk The order in which applications are listed is not significant An asterisk in the In Use column indicates that an application has been selected for loading onto one or more slots See Section 2 3 If there are more than 10 installed applications you can move backwards and forwards through all the entries by using the First Previous and Next page options 5 5 2 Installation The Install command allows you to install a new application onto the PSE from a distribution disk When you select the Install command you will be prompted to insert a distribution disk into drive B and type RETURN You then install the application by selecting PF1 or abort the installation by selecting the PF3 or PF4 key If PF1 is entered the application is installed and the list of installed applications is updated on the screen The Node Manager copies all the files needed to support the application from the application distribution disk to the system disk The files copied will be the application s load and database files The database files applic data and novid data on the system disk will be updated to hold details of the new application Different applications builds can share load files If an application or file is found to exist already on the system disk then the operation will be abor
49. specify the appropriate build of the Xpress Kernel software kerU03X L for an 8325 card and UPM3 and UPM4 The Install and Delete commands will not manipulate the Xpress Kernel load files The Install Delete and Expand commands use the database files on the system disk and when appropriate on the application distribution disk as their source of information about the files which must be operated on The Install Delete and Expand commands are as much as possible transaction oriented i e all disk accesses will be performed at a single point in an operation If an access fails the file system will be restored toa consistent state 5 5 6 Application Specific Files The Install Delete and Expand commands do not manipulate types of file specific to an application e g configuration files You must use the existing disk utilities to copy such files to the Xpress system disk delete them from the system disk and list them The NMC will be unable to request the PSE to do Configuration U pload Download of an application s configuration files because the PSE will not be aware of such files X890 304751 Issue 1 5 13 Rev 0 5 6 Print Utilities The following print reports are available Routing configuration report including hunt and trunk groups public data network gateway addresses the inter node routing table local to global CUG index map Address Analysis Table DNIC Barring Table X 25 X 75 or trunk port configuration report
50. supported at ports configured with a V 24 physical interface A special cable is needed to connect the port to the modem part no X890 406311 The PSE will generate the remote loopback RLB signal on circuit number 140 pin number 21 The PSE will generate the local loopback LLB signal on circuit number 141 pin number 18 The PSE will receive the Test Indicator signal on circuit number 142 pin number 25 X890 304751 Issue 1 F 4 Rev 0 Appendix G The Broadcast System G 1 Introduction The Xpress PSE operating software now offers a data broadcast facility known as the ABS Asynchronous Broadcast Service The ABS isa distributed facility accessed via the Virtual DTE mechanism The ABS broadcasts data received from a single source to multiple destinations It does this by establishing an X 25 call to the data source or host and then making multiple copies of every data packet received over this call and sending one copy to each user or client currently connected to the service The ABS is implemented as a co operative collection of servers i e clients must make an X 25 call into one of the ABS servers in the system in order to pick up the required broadcast data All other users on the network are unaffected by the presence of the ABS Each ABS server can support up to 32 clients at a time To build broadcast systems with more than 32 clients several servers can be combined in a distributed hierarchy Thus there
51. system by allowing the system manager to define the users that are allowed logon rights tothe PSE The system holds a pre defined list of valid user logon names and associated passwords Each user name has an access profile or user type identified with it which specifies the access rights for that user Up to eight user types can be allowed logon rights to use the system The manager functions have been logically divided into several distinct areas e g routing specification configuration Profiles can be set up for each user to define the user s access permissions read only write or no access to these functions If at any time whilst using the system you get either message No access or Read only access this is because your access profile prohibits the operation you are trying to perform Only the system manager or a very small number of users should have access to utilities which can change the PSE s configuration Every user on the system can have access to the Change User Password screen X890 304751 Issue 1 5 1 Rev 0 5 1 1 Change User Password You must first enter your current password correctly If you fail to do so in three attempts you must start again Passwords may be up to 16 and preferably more than 4 characters long and may use any of the characters a toz AtoZ and 0 to9 Anything else will be rejected with the message Illegal password entered To change someone else s password only the system manager should b
52. take approximately 30 minutes to transfer all the files over a 19K 2 trunk link 6 Logout of node 1 Login to node 30 and list the contents of drive b to check that all the files are present 8 Fromthe utilities Disk Utilities File Copy screen copy all the software files from drive b to drivea The from file will be a aLt once again but the to file will be just b 9 Wait for the operation to complete It will take approximately five minutes 10 Restart the node using the Configuration Node configuration Edit node configuration Node State restart option 11 You will get cleared 12 Give the node time to reboot and then login to node 30 and check that it is functioning correctly L 3 3 Remote Software Version Download 8425 8525 Before you attempt any of the following operations you must have backups available of all the disks on all of the nodes that you will be accessing This procedure can be used to update the software of an 8425 or 8525 node remotely Assume the following Node number 1 is the central node and is running Version 8 2 A remote node number 30 is running Version 8 1 and needs to be upgraded to Version 8 2 All nodes are 8525s A TGate application is installed The operator is working from the Mini Pad at node 1 An 8425 or 8525 supports a number of X 25 cards XIM UPM XIM UPM3 SPXIM UPM3 Each card requires a load file If a node contains all of the X890 304751 Issue 1 L 5 Rev 0 various types
53. the DCE and DTE These frames consist of address control and check sequence fields which may enclose a packet of user data Frames may hold any data patterns and are delimited by flags A flag is a unique bit pattern and to prevent the pattern occurring within a frame the technique of bit stuffing is used i e the transmitter inserts zero bits into the data stream to avoid the occurrence of flags inside a frame The receiver carries out the reverse process to remove the extra bits from the data stream and thus restore the original data Span of Zero Insertion Span of CRC Flag Variable Length Frame check Flag 01111110 Information field Sequence 01111110 8 bits 8 bits 8 bits 1 Multiple of 8 bits FCS or CRC Max 32768 lt gt Level2Frame User data 16bits Header Figure A 2 HDLC Frame Structure X890 304751 Issue 1 A 4 Rev 0 The address field of frames holds little information because as Level 2 operates over a point to point link only the direction of transfer need be identified The control field identifies the type of frame Different types of frames are used to establish and maintain LAPB provide flow control recover from errors and to carry packets This last type of frame is called the information frame or frame Each frame carries one and only one packet A 2 2 3 Level 3 The Packet Layer This describes the exchange of packets between the DTE and the
54. the node have been set as required will be displayed and you will reenter the Node Configuration Menu Enter the Configuration Node Configuration detailed link status display screen and verify that the ports are Online You can also check the state of the X 25 level 2 Data Link which will either be Up or Down If both ports are Up it means that the data links are waiting to carry packet level calls If either link is down refer to Section 3 7 4 below Now try the call Assume that there is a suitable asynchronous terminal connected to port 1 of PAD 1 and port 6 of PAD 2 and that both PADs support the X 28 command set On PAD I s terminal type CON 11000010002006 RETURN This should cause both terminals to display COM and they should be connected together If not refer to Section 3 7 5 below Type a few lines on both terminals to ensure that traffic flows in both directions You may have to hit RETURN before anything will be sent To clear the call type CTRL P CLR PAD 1 s terminal should display CLR CONF and PAD 2 s terminal should display CLR DTE or similar To make a call in the opposite direction follow the above procedure on PAD 2 s terminal using an address of 11000010001001 Clear the call in the same way See Section 3 5 3 for an explanation of Xpress internal addressing X890 304751 Issue 1 3 34 Rev 0 3 6 2 Example 2 Application Port Configuration This second example follows on from the first example It shows
55. the order physical level for all ports data link level for all ports etc rather than 1st port physical level 1st port data link level etc Enter the physical level screen and enter PAD 1 s logical port number 0001 Thedefault configuration must be changed as PAD 1 does not supply a line clock Type s RETURN andi RETURN to set the port to use the internal Baud Rate Generator and thus supply the line clock Type c RETURN and 4 RETURN to set the baud rate to 9600 bps Press PF1 to submit the screen Thelnterface Type and Transmit Flag Insertion parameters are compatible with the PAD 1 and PAD 2 so we don t need to change them X890 304751 Issue 1 3 32 Rev 0 12 13 14 15 16 17 The message Configuration change completed successfully will be displayed Type r RETURN or k RETURN and type PAD 2 s logical port number 0002 PAD 2 provides the line clock at 19200 bps If you typed k at the start of this step the screen will have been left with port 0001 s setting so type s RETURN and e RETURN to set port 0002 to use the external clock If you typed r RETURN the screen will have been given default settings so we need not change the clock source Type c RETURN and 5 RETURN to set the clock rate to 19200 bps This is optional but should be done to ensure maximum performance Press PF1 to submit the screen and PF3 to return tothe X 25 Port Configuration Menu Both
56. to ensure that calls made in both directions across it are rerouted X890 304751 Issue 1 4 8 Rev 0 4 3 2 Trunks over Frame Relay One of the most important uses for the Xpress frame relay interface is to allow multiple Xpress trunks to be carried over a single physical interface into a frame relay network This allows the use of a high speed frame relay network such as the Cray FPX2000 to provide a backbone for groups of Xpress nodes acting as frame relay concentrators for existing equipment which is not frame relay capable Xpress trunks are multiplexed over frame relay networks using the same frame relay virtual physical port mechanism as that used by X 25 X 75 ports and may in fact be freely mixed with such ports as required E g it is possible to multiplex two trunk ports and four X 25 ports over a single frame relay physical interface for connection to two remote Xpress nodes and four remote hosts The following example works through the configuration of two Xpress trunks on an 8425 8525 SP XIM multiplexed over a single frame relay interface connected to a Cray FPX2000 network X890 304751 Issue 1 4 9 Rev 0 Logical trunk from node 1 to 2 on frame relay virtual physical link 0 7 0 T0070 multiplexed over physical link 0 7 0 to FPX and identified by local DLC 16 _ a ae Physical frame relay link connected to slot 7 port 0 EPX2000 Nodes Node 1 mee baat EON eet Paar ae e Logical trunk from node
57. to the attached device If REJ ECT then any call request containing a NUI will be cleared The default is FORWARD RPOA Subscription If a DNIC is entered in this field then any calls received from the attached device which do not hold an RPOA will have the DNIC inserted as the RPOA Selection The call will be routed by the RPOA instead of its called address Default is no DNIC X890 304751 Issue 1 3 21 Rev 0 e DNIC of Local RPOA This specifies the DNIC of the attached network If the first RPOA Selection in a call request from a remote device matches this DNIC then the RPOA Selection entry will be removed before forwarding to the attached network Default is no DNIC e TNIC Suppression X 75 only If NO is selected then this parameter specifies that Xpress will add its Internetworking DNIC tothe list of Transit Network Identification Codes TNIC before forwarding it to the attached network If YES then Xpress will not add its Internetworking DNIC tothe TNIC list in the call and clearing request packets The default is NO e CNIC Suppression X 75 only If NO is selected then this parameter specifies that the Clearing Network Identification Code CNIC is to be forwarded to the attached network If YES then the CNIC will be suppressed from the clearing packet The default is NO 3 4 6 Congestion Monitoring It is unusual for Congestion Monitoring and Control to be required to operate on this type of port i e an X 25 X 75 port
58. up using Type specification A profile defines the user s access rights to the system management functions Comments can be further information to help identify a user e g finance department 5 1 3 2 Delete User This screen enables the operator to remove a user from the system The user is identified by the username Once deleted the user may no longer logon to the system For security purposes the user wizard cannot be deleted 5 1 3 3 Edit User Attributes This screen enables the operator to change the attributes of a user The password comments and the user type may be changed Use the Access utilities User access specification Change password screen to change your own password For security purposes the attributes of the user wizard may not be altered Logon as wizard and use the Change password screen to change the wizard s password 5 1 3 4 List Users This screen displays all users that may logon to the system For each user the following details are displayed X890 304751 Issue 1 5 3 Rev 0 Name the user s name Type the user s access level Comment a short description of the user If the list of user names exceeds one screen full use Next page and Previous page to view the complete list 5 1 4 Initial Users When the equipment first arrives the PSE is programmed with four users These users have varying access to the screens They are able to read all the screens in the system Type Username Password
59. which specifies a standard international interface between X 25 PDNs for the purpose of forwarding X 25 calls over two or more PDNs X 75 provides facilities which network administrators can use as the basis for a Improving security by barring X 25 calls which originate from or are destined for specified networks see Section 3 9 5 about the DNIC Barring Table b Improving routing management by preventing loops through transit networks c Collecting statistics by recording the transit networks via which a call is set up and also any network responsible for clearing an established call d Billing because X 75 carries a unique identification of an X 25 caller X 75 also aids X 25 call set ups because it allows most X 25 User Facilities to be carried transparently over transit networks and X 75 interfaces X890 304751 Issue 1 A 15 Rev 0 Xpress supports X 75 as a method for connecting an Xpress network with one or more other Xpress networks and or other types of private and public X 25 networks Xpress supports X 75 1980 as specified by the CCITT in the Yellow Book and X 75 1984 as specified in the Red Book X 75 1984 is much more like X 25 1984 88 than X 75 1980 is The PSE allows ports to be configured for X 75 1980 or X 75 1984 Unless otherwise stated the text refers to X 75 1984 A 2 7 2 How the X 75 Protocol Works X 25 defines the protocol between a DTE and DCE TheX 75 Recommendation de
60. with X 25 card a ISDN TA ISDN ISDN TA Figure K 1 Example of Dial up Link and Trunk Usage X890 304751 Issue 1 K 1 Rev 0 In Figure K 1 the PC is connected to node 1 via a V 24 dial up modem link and node 1 connected to node 3 via a dedicated circuit backed up by a V 11 dial up ISDN TA trunk In normal operation with no X 25 calls present on either of the dial up links DTR will be low on the modem ports and Control low on the ISDN TA trunk ports hence neither physical link will be active and all the dial up X 25 ports and trunks will be down If the PC makes an X 25 call to the host it raises DTR to its local modem which then dials the phone number of the modem attached to node 1 This modem starts pulsing Ring Indicator and the node will raise DTR on the dial up port to instruct the modem to answer the phone call The modem answers trains up and X 25 levels 2 and 3 come up between the PC and node 1 The X 25 call is then forwarded across the link from the PC to node 1 where it will be onward routed across the dedicated trunk to node 3 and on to the host Assuming the secondary route from node 1 to node 3 is via the dial up trunk rather than via node 2 and that the primary trunk fails the user s X 25 call will be internally cleared back to the entry port and will then re establish via the dial up trunk The routing software will note that the trunk is a dial up trunk and will hold onto the re establishment call request while it
61. 0 09 00 Time mins Figure M 3 Utilisation Graph for Node 3 T31 Priority 1 calls Priority 2 calls Priority 3 calls Time mins Table M 1 Profile of Calls Present on Node 3 T31 f With the restoral of the terminal session utilisation starts to rise however the workstation clears its call at 6 minutes 45 s which causes a reduction towards 70 g This is still above the refuse threshold for the printer jobs so they are unable to return to the primary trunk when the Auto Reroute cuts in at 9 minutes on the secondary X890 304751 Issue 1 M 13 Rev 0 M 4 Summary Points Some points to remember when configuring Congestion Monitoring and Control on trunks Utilisation levels can be observed on the status screen Configuration Node Status Display Detailed Link Circuit Display to help in assessing whether there are congestion problems on a port and for determining Suitable threshold values The utilisation of a trunk or link is separate for each direction of traffic The Congestion Monitoring feature considers utilisation only in the outward direction from the port Therefore to find out the utilisation in the other direction it has to be inspected at the remote port via the Manager of the next hop node Remember that Congestion Monitoring and Control is generally useful only on ports which have at least one alternative route Such as a secondary or tertiary next hop trunk port to take up any calls that get refused or bump
62. 0001 you don t have to type leading zeros Type RETURN at the end Type ph RETURN Type 1 RETURN 0 RETURN to select slot 1 link 0 Type po RETURN PAD Number 1 RETURN to set the port description Press PF1 to submit the screen The message Logical port created successfully will be displayed If it isn t refer to Section 3 7 3 below Type r RETURN to repeat the action Type in the logical port number for PAD 2 Use logical port 0002 X890 304751 Issue 1 3 31 Rev 0 Note that the settings for Bay Slot and Link have been left alone but that the port description has been set to Port number 2 Type ph RETURN Type 4 RETURN 2 RETURN to select slot 4 link 2 Type po RETURN The other PAD RETURN to set the port description Press PF1 to submit the screen Again Logical port created successfully will be displayed Refer to Section 3 7 3 if it isn t 9 Toset the X 25 configuration of the two logical ports press PF3 to return tothe Logical Port Allocation Menu Enter the logical physical port display screen to check that all is well This screen displays the allocated logical port numbers together with the Bay Slot and Link numbers to which they are assigned and their port descriptions 10 Press PF3 twice to return tothe Configuration menu and enter the Port Configuration X 25 Port Configuration Menu 11 Because of the way the menus work it is always easier to configure ports in
63. 25 and X 75 protocols allow each physical port to be logically divided into 4096 logical divisions or channels Each of these channels can carry a single call or virtual circuit When a call is set up it is assigned to a free Logical Channel and all packets belonging to that call will be identified by the number of that Logical Channel The Logical Channel is freed for re use when the call is cleared There are four groups of LCNs Permanent Virtual Circuit group Incoming Call Only group Two way Call group Outgoing Call Only group These must be set up to match the configuration of LCN groups inthe connected device A group is defined by the lowest and highest LCN within it The valid range for LCNs is 0 4095 An empty group is denoted by setting its low and high LCNs to none There may be gaps between groups but they must not overlap Only two way LCNs can be configured for Application Links The UPM1 and UPM2 XIM cards can support up to 256 LCNs The UPM3 XIM and SP XIM cards can support up to 512 LCNs The UPM4 based SP XIM cards can support up to 1024 LCNs Only up to 256 X890 304751 Issue 1 3 16 Rev 0 LCUNS Can be Supported tor Application LINKS Irrespective or Tne card types Examples of valid LCN Ranges Group LowLCN High LCN PVC none none Incoming none none Two Way 1024 1031 Outgoing none none This arrangement is the default for XIMs It has 8 Two Way LCNs asit is commonly used with 8 port PADs
64. 4 bytes exchanged during call number of segments sent by DTE X890 304751 Issue 1 5 17 Rev 0 5 9 Billing Billing information should not be confused with Charging information which is explained in Section 5 8 Billing information is generated by the called and calling ports when an established X 25 X 75 call clears down Each billing record contains among other things the called and calling X 121 addresses the number of bytes of data transferred the duration of the call etc The format of a billing record is given in Appendix C The billing record can therefore be used by the system manager to analyse the usage of the PSE and also as a basis for customer charges A billing destination i e any X 25 port capable of receive the billing records can be specified by means of its X 121 address The billing destination will normally be elsewhere in the network If the destination is not specified billing records will be discarded A number of configuration options are provided e g for restricting billing to include only successfully connected calls This is so that the amount of billing information forwarded across the network can be minimised if required 5 9 1 How Billing Works The billing system works in two parts 1 Collection of the billing record from the X 25 X 75 software on the XI Ms 2 Forwarding it to the billing destination The record is stamped with the time in HH MM SS DD MM YY format The Node Manager colle
65. 9 X890 304751 Issue 1 G 12 Rev 0 Node 3 Match Address Internal Address 1100 003 9049 995 1100 003 9079 005 1100 003 9079 995 1100 002 Map server 4 s host call to server 7 and server 7 s host call to node 2 Node 4 Match Address Internal Address 1100 005 9019 995 1100 002 Map node 5 server 1 s host call through to node 2 Node 5 Match Address Internal Address 1100 005 9019 995 1100 002 1100 005 9029 995 1100 002 9019 005 Map server 1 s host call to node 2 and server 2 s host call to server 1 X890 304751 Issue 1 G 13 Rev 0 G 4 Providing More than One Broadcast Service It is possible to arrange for two or more completely separate collections of servers to broadcast data from multiple hosts to multiple client populations The example in Figure G 5 shows the same network as that used in Figure G 4 but with two hosts broadcasting data to their own population of clients Obviously the clients in either population could decide to pick up the broadcast data from the other host simply by calling a different server X890 304751 Issue 1 G 14 Rev 0 Host These 4 PADs can support up to 31 clients I Port 32 aas Tor Node 4 N N L DN I I N I N I N Port 27 NS Ports 1 0 1 3 l Node 4 Node 2 Ports 20 23 L C c ee Node 3 DO nn n Pon Eee C 70 733 NC es Doo Ports 40 43 These 4 PADs can support E up to 31 clients These 4 PADs can support up to 32 clients L C Key
66. ANSI standards Management of the PSE is via a VDU based Node Manager using a user friendly menu system The Node Manager can be accessed from a local terminal or remotely from another Xpress PSE In addition the PSEs can be controlled from a Cray Network Management Centre NMC or any compatible network terminal An NMC provides the ability to logon transparently to the PSE Node Manager upload and download configurations and to capture and report PSE events In addition to packet switching facilities the Xpress PSEs also provide support for ancillary software applications running within the Xpress hardware software environment For information about additional Xpress PSE applications refer tothe appropriate user documentation supplied with each application This manual describes only how to install applications onto a PSE and manage them The Cray Access Control Server ACS provides a network security service for an Xpress network X890 304751 Issue 1 1 1 Rev 0 X890 304751 Issue 1 1 2 Rev 0 2 Getting Started 2 1 Powering Up the PSE This section explains how to power up the system from cold and the resulting sequence of events Any error situations that may arise are also explained along with remedial action The disk file structure and naming conventions are also explained 2 1 1 Booting Up the System 1 Place the system disk in drive A Placethe dump disk in drive B If you put the disks into the wrong drives the syst
67. Billing records will be forwarded to the billing destination for all types of call User calls Billing records will be forwarded only for user calls This excludes billing for calls to or from the Network Management Centre Specifically calls to or from an address that starts with 999 Successful user calls Billing records will be forwarded only for user calls as defined above which got connected Off No billing records will be forwarded they will all be discarded Except when Billing is set to Off the Node Manager will attempt to set up a call to the specified billing destination If it is unsuccessful an event will be generated Check that the address specified is valid It may also be that no path exists to the billing destination because an intermediate link is down The Node Manager will in any case re attempt the call set up every 4 minutes until successful X890 304751 Issue 1 5 19 Rev 0 5 10 Statistics The Statistics utility can be configured by the user to collect statistics for selected PSE components at selected intervals Statistics reports can be generated on the printer at specified regular intervals They can be examined per port on demand via the Node M anager screens Statistics are collected regularly for any specified X 25 X 75 or trunk ports any specified module ACM UPM pair or for the Intra N ode Communications Subsystem INCS The options available are described in the fo
68. Call Request packet are processed in the following order as it passes through an Xpress network Entry Port Incoming Call Address Translation I CAT Called destination and Calling source Address If translation has occurred then the address will have been overwritten Note A call request containing RPOA will also have Address Translation done DNIC Barring Table if Entry Port is an X 75 Gateway X890 304751 Issue 1 4 14 Rev 0 Per node Internetworking DNIC used to strip RPOA if aRPOA Subscription remains in the call request then Route to network specified by RPOA The PDN Gateway Table see Section 3 9 is examined for the node having a gateway to that network there is no RPOA so analyse the called address using the Address Analysis Table AAT This gives a match address to which the PSE will route the call if the match address starts with the Xpress DNIC 1100 then the destination node number is encoded in the address itself 1100 001 0020 001 destination node number the match address does not start with the Xpress DNIC The DNIC in the match address may be for another X 25 network The PDN Gateway Table see Section 3 9 is examined for the node having a gateway to that network endif endif NOTE The addresses in the call request packet are left unchanged Outgoing Call Address Translation OCAT Called destination and Calling source Address If translation has occurred then the ad
69. DCE Level 3 transfers packets by making use of the services provided by Level 2 Level 3 of X 25 is also covered by ISO OSI standards such as International Standard 8878 and International Standard 8208 1984 version of X 25 Level 3 manages logical channels and provides SVCs and PVCs It multiplexes virtual circuits over a link sothat a DTE may have many concurrent connections to many other DTEs via the Packet Switching Network Level 3 provides flow control on a per virtual circuit basis It also provides User Facilities such as Reverse Charging delivery confirmation of packets etc These User Facilities are specified in Section 3 4 5 Level 3 allows certain packets to be qualified this feature facilitates the transport of other protocols such as for PADs or IBM s SNA across X 25 X890 304751 Issue 1 A 5 Rev 0 Level 3 handles packets of information Each packet consists of a header and user information as shown in Figure A 3 GFI and LCN Packet Variable length information field LCGN type 8 bits 8 bits 8 bits Multiple of 8 bits usually 128 bytes Call request Facility requests and addresses Data packets anything PAD control Protocol identifier X 29 Figure A 3 Level 3 Packet Structure The packet header contains information about the format of the packet It also identifies the virtual circuit with which the packet is associated by means of the Logical Channel Group Number LCGN and Logical Channel Number
70. ECN FRAD Gateway HDLC Hz kbps ICAT INCS ISDN ISO LAPB LAPD LCI LCN LCGN Data Circuit terminating Equipment Data Concentrating Exchange DNIC Barring Table Discard Eligibility bit A disk holding an application s database files and load files Data Link Connection Identifier Data Network Identification Code Data Terminal Equipment F rame Check Sequence Forward Explicit Congestion Notification Frame Relay Access Device This is an X 25 port which is used to interface Xpress nodes toa PSPDN or PSPvtDN High level Data Link Control Hertz Kilobits per second Incoming Called calling Address Translation ntra Node Communications System This allows UPMs to communicate with each other within a PSE Integrated Services Digital Network International Standards Organisation Link Access Procedure Balanced Link Access Procedure for ISDN D channel Logical Channel Identifier this comprises the LCGN and LCN Logical Channel Number Logical Channel Group Number X890 304751 Issue 1 H 2 Rev 0 LMI LPN M bit MMI Module ms Native application NMC NMS Node OCAT OSI PAD PDN PSE PSPvtDN PSPDN PSS PSU PVC Q bit RAM ROM RPOA SAC SAM Local Management Interface Logical Port Number X 25 Level 3 More data indicator Man Machine Interface A UPM or an ACM millisecond s Xpress standard software such as the Node M anager or X 25 software which does not use the Xpress Kernel Network Manag
71. Loager Cards J 1 3 Imported Applications J 2 Network Architecture J 2 1 X 25 J 2 2 Network Management Service J 2 3 Network Connectionless Service J 2 4 Node Connectionless Service J 3 Hardware Architecture J 4 Software Architecture J 5 Application Programming Interface API J 5 1 Overview J 5 2 Applications Environment J 5 3 Management Services K Dial up Ports K 1 Overview K 2 Operation and Signalling K 2 1 General K 2 2 V 24 Interface Circuits K 2 3 V 111nteface Circuits L Remote Software Download L l Version 8 Features L 1 1 Remote File Operations L 1 2 Enhanced Pattern Matching L 13 Self Extracting Compressed Load Files L 14 Move Command L 1 5 Node Restart L 2 Security Considerations L 2 1 File Corruption L 2 2 Security Violations L 3 Example Operations L 3 1 Configuration File Backups L 3 2 Remote Software Version Download 8325 L 3 3 Remote Software Version Download 8425 8525 L 3 4 Remote Installation of Applications L 3 5 Points to Beware Of L 4 Software Licensing M Congestion Monitoring and Control M 1 Introduction X890 304751 Issue 1 0 11 J 1 J 1 J 2 J 2 J 2 J 2 J 2 J 3 J 3 J 6 J 6 J 6 J 7 K 1 K 1 K 3 K 3 K 3 K 4 L 1 L 1 L 1 L 1 L 2 L 2 L 2 L 3 L 3 L 4 L 4 L 5 L 7 L 9 M 1 M 1 Rev 0 M 2 22 Z NF O A 1 A 2 A 3 A 4 A 5 A 6 A 7 A 8 A 9 D 1 Parameters to be Configured M 2 1 Configuring the Congested Trunk Port M 2 2 Configuring X 25 X 75 Li
72. Network Address must be a valid 8000 Series address Notes ThePSE supports the basic Call Redirection Deflection service which X 25 defines as one Redirection Deflection per call set up Call Redirection only works when the port is On Line Calls may be redirected deflected to a foreign network This does not follow the CCITT recommendation Call redirection will occur on ports within a hunt group Network Data Integrity If this parameter is enabled the PSE will protect against loss of data in transit within the network when calls that have originated from this port are re routed or internally reset Calls using D bits are automatically provided with network data integrity By default this parameter is disabled Local NUI Selection X 25 only Network User Identity NUI Selection is an X 25 1988 facility This parameter specifies how the PSE handles an NUI in the call request received from the attached device If ALLOWED is selected then call requests may contain the NUI If DISABLED then call requests with NUI will be cleared If REQUIRED then call requests must contain NUI The default is ALLOWED Remote NUI Selection X 25 only This parameter specifies how the PSE handles an NUI in the call request received from the network i e a remote device If FORWARD is selected then the call request will be forwarded to the attached device If REMOVE then the NUI will be removed from the call request before forwarding
73. PADs have configurations which are compatible with the X 25 level 2 default configuration so you don t actually need to enter the data link level Screen Enter the network level screen and type in PAD 1 s logical port number 0001 Type t RETURN to set up the Two Way LCN range Type I RETURN 1024 RETURN to set up the lower LCN Type u RETURN 1031 RETURN to set up the upper LCN Press PF2 to escape from the boundary change prompt The rest of the configuration is correct so press PF1 to submit the screen The message Configuration change completed successfully will be displayed PAD 2 s network level configuration is identical to that of PAD 1 so type k RETURN 2 RETURN PF1 to configure its port s network level Press PF3 to return tothe X 25 Port Configuration menu To keep things simple assume that neither PAD requires changes to the user facilities screen although you may care to look at it and to try switching flow control parameter negotiation to YES X890 304751 Issue 1 3 33 Rev 0 18 19 20 Either usethe change state of port screen to bring the two ports Online or use the alternative method explained below Press PF4 to return to the main menu Enter the Configuration Node Configuration change state of all ports on node screen This screen should show 2 ports in Out Of Service state Type on RETURN PF 1 to bring the ports Online After a short pause the message The port states on
74. Physical Level Statistics screen The PSE may be configured to monitor the Test I ndicator signal raised by a V 54 modem This signal is raised by the modem to acknowledge that it has entered loopback If being monitored the PSE will raise an event whenever the signal changes state This configuration is saved on disk and so will be remembered if the PSE is restarted X890 304751 Issue 1 F 2 Rev 0 F 3 Test Pattern Generator A test pattern generator at every PSE port allows test messages to be transmitted to the local modem If used in conjunction with a modem test loop it will check the looped back data for errors Thetest pattern generator can be enabled independently of the test loops and so can still be used if a loopback has been established manually When activated the test pattern generator starts sending frames at regular intervals and awaits receipt of similar frames When the first good frame is received the PSE raises an event to indicate that the test pattern generator has achieved synchronisation Whenever no good frames have been received for 30 seconds the PSE raises an event to indicate that the test pattern generator has lost synchronisation The current state of the test pattern generator and number of frames transmitted and good bad frames received is available via the node manager X 25 Port Trunk Physical Level Statistics screen X890 304751 Issue 1 F 3 Rev 0 F 4 Signals and Cables Modem test loops are only
75. Priority Class for each call arriving from a pre Version 9 part of the network X890 304751 Issue 1 M 8 Rev 0 M 3 Using Congestion Monitoring and Control M 3 1 Description of the Example Network This section illustrates the use of Congestion Monitoring and Control for an imaginary network shown in Figure M 1 The primary route between Nodes 1 and 3 is backed up by the secondary route via Node 2 NODE 2 T21 T22 i 60 R F M P N Ai ieee a pak Ta NOE A S L C S EE 50 40 HOST D COMPUTER z LI PRINTER WORKSTATION Figure M 1 Example Small Network with Trunk Congestion Occurring There are three types of device attached to Node 1 that are using the network Priority 1 a workstation requiring a very fast response with minimum delay by the network Priority 2 7simpleterminals having interactive sessions via asynchronous PADs which require a generally quick response time X890 304751 Issue 1 M 9 Rev 0 Priority 3 low urgency spooled printer jobs to 2 printers The utilisation of the primary trunk between Node 1 and Node 3 trunk has been observed in the outgoing direction from the central site by inspection of UTLS on Configuration Node Status Display Detailed Link Circuit Display at Node 3 T31 Whenever utilisation exceeds 75 the workstation user is experiencing unacceptable application timeout problems The terminal users response times noticeably worsen at this level but remain acceptable until uti
76. RETURN just press RETURN immediately followed by a PF1 submit After the installation is complete you should see that the TGate application is now available again Check that you see the entry 1 Tgate Telnet Triple X Gateway Co res on the screen you might have to issue a Next page command if multiple applications etc are available Finally all that remains is torestart the node using the Configuration Node configuration Node start restart option You will get cleared Give the node time to reboot and then login to node 30 and check that it is functioning correctly L 3 4 Remote Installation of Applications The remote installation of applications is covered under steps 10 to 15 of the previous Section L 3 3 The only difference is that under 15 you just restart the appropriate card rather than restart the entire node X890 304751 Issue 1 L 7 Rev 0 L 3 5 Points to Beware Of Boot File There is a restriction placed on the boot file of an Xpress format disk The file v8boot or v8xboot must be the first file on a disk So after formatting a new disk assuming you want the disk to be bootable you must copy the boot file from another disk immediately before you store any other files If you are downloading intermediate versions of software i e V8 1 to V8 2 then there is no problem since the boot filename stays the same But if you are upgrading to a completely new version you must renamethe old boot filename to that of t
77. S and WizzNet Incoming PSS calls are subject to ACS control 4 The master ACS does not directly support any users but controls the slaves and acts as backup to them both in the case of failure 5 Weshall not concern ourselves with the configuration of X 25 levels 1 and 2 or the basic network part of level 3 but shall concentrate on the level 3 user facilities and network addressing E 4 1 ACS Port Configuration On all ACS ports node 1 port 0002 node 2 port 0002 node 3 port 0003 the Call Deflection and Call Deflection in Data Transfer facilities must be set to YES This allows the ACSs to transfer user calls The port type must be set to 1980 CCITT In order to provide ACS resiliency we assign the ACSs a Public Data Network PDN Data Network Id Code DNIC which is otherwise unused in the network e g DNIC 9998 This allows us to use the standard PDN Gateway support to provide multiple gateways into each ACS which in fact can be used to route calls to the requisite backup e g Master ACS in the case of ACS failure X890 304751 Issue 1 E 6 Rev 0 Note that we have chosen DNIC 9998 as 9999 is reserved for the NMC We now Set up each node s view of what the DNIC 9998 means Node 1 This node has no ACS users but has the master ACS connected to it Primary Gateway Node 1 port 0002 Secondary Gateway 9 Tertiary Gateway Node 2 This node supports community A who are served by ACS slave 1
78. Series 8000 Xpress PSE User Guide Software Version 9 Case Communications Ltd 1997 X890 304751 Issue 1 0 1 Rev 1 Unit 15 Riverside Business Centre Victoria Street High Wycombe Bucks HP11 2LT Web www casecomms com Email sales casecomms com Tel UK 08700 263 740 Tel International 44 0 1494 833 740 Fax UK 08700 263 741 Fax International 44 0 1494 833 741 STATUTORY NOTICES For the statutory notices relevant to each of the Series 8000 Xpress PSEs covered by this manual please refer to the following installation guides 8325 Installation Guide X890 303251 8400 Installation Guide X890 301451 8425 Installation Guide X890 302551 8500 Installation Guide X890 302651 8525 Installation Guide X890 302151 These manuals cover the relevant system with Version 9 operating software Case Communications Ltd declare that this product conforms with the protection requirements of Council Directive 89 336 EEC on the approximation of the laws of the member states relating to electromagnetic protection This equipment has been tested using shielded cables supplied by Case Communications Ltd These cables or equivalents must be used to ensure compliance with this declaration All PCB assemblies contain Electrostatic Sensitive Devices ESDs which may be permanently damaged if incorrectly handled This equipment must be handled in accordance with BS5783 code of practice for the handling of electrostatic sensitive devices
79. The screen may be refreshed by typing F to redraw the first page X890 304751 Issue 1 5 26 Rev 0 5 11 3 Display Link Circuits This screen displays details of the calls in progress on a specific logical port The first line of the display provides details of the specified logical port the configured state totals of PVC and SVC calls and the number of calls of each priority class For each call the following is displayed LCN logical channel number Called translated called address Calling translated calling address Status current state of the call Via the corresponding logical port on which this call leaves the PSE Pri Priority Class The call status can indicate the following values CIP connection in progress DATATFR call is connected DIP disconnection in progress The Trace option on the command line allows an individual virtual circuit to be traced When a call is traced a Trace Event is generated at every XIM crossed by the virtual circuit Trace events are logged on the system printer and collected by the Network Management Centre for analysis and display 5 11 4 Summary Link Display This screen displays the X 25 X 75 protocol state of every physical port in the bay Up or Down Ports that have not been allocated a logical port number are displayed as Ports on a module which is not Operational are displayed as X890 304751 Issue 1 5 27 Rev 0 X890 304751 Issue 1 5 28 Rev 0 6 Diagnostics and Erro
80. The valid range for this parameter is 1 to 10 monitored events with a default of 5 e R bit support The R bit is a proprietary mechanism used by the Cray FPX 2000 frame relay interface to implement explicit per DLCI congestion notification in full status messages This parameter should always be set to yes when connecting to F PX2000 frame relay interfaces or no when connecting to other frame relay services The default is no e Bidirectional Procedure Currently Xpress ports do not support the optional bidirectional LMI procedure where the network can status poll Xpress and this parameter should be set to no which is the default X890 304751 Issue 1 3 13 Rev 0 e Polling Verification Timer T392 This timer is not currently used as it is required only to support the bidirectional LMI procedure Consequently the configured value of T 392 is ignored 3 4 3 Data Link Level Parameters These parameters are used to control aspects of the operation of the X 25 X 75 Level 2 LAPB data link The Level 2 software is sufficiently flexible in operation to start up and run automatically with any compatible LAPB The configurable data link level parameters can normally be left set to their default values The first nine parameters apply to all logical ports other than those assigned to internal Xpress applications The last two parameters are configurable only for logical ports which are mapped to frame relay virtual physical ports
81. X 25 Recommendation is an international standard produced by the CCITT In 1980 the CCITT produced the Yellow Book version which specified a new link level protocol LAPB and standards for PADs Most of today s networks conform to the 1980 version of X 25 In 1984 the CCITT published the Red Book version of X 25 which adds various enhancements and support for Open Systems Interconnection OSI In 1988 the CCITT published the Blue Book version of X 25 which adds further enhancements The PSE allows ports to be configured to offer the facilities of X 25 1980 X 25 1984 or X 25 1988 In this appendix the text indicates where a 1988 feature is being described A 2 1 Other Standards Relevant to X 25 X 25 is one of anumber of CCITT Recommendations which apply to Packet Switching Other related standards are X 1 Signalling rates X 2 List of user facilities X 3 X 28 X 29 Standards relating toPADs These three standards are commonly referred to as Triple X X 32 Dialup X 25 X 75 Procedures for links between two X 25 PDNs See Section A 9 X 121 Numbering plan for PDNs X 21 and X 21bis Procedures for the electrical interface X890 304751 Issue 1 A 2 Rev 0 System 1 System 2 Layer Application Application 7 Presentation Presentation 6 Session Network relay Session 5 Transport Packet Switch Transport 4 Network Network Network 3 Link Link Link 2 Physical Physical Physical 1 wires wires
82. XIM_ Two protocol ports via physical ports 0 to 1 plus two FR virtual physical ports 2 to 3 multiplexed over FR physical port 2 1 X 25 3 FR UPM3 XIM_ Protocol port via physical port 0 plus three FR virtual ports 1 to 3 multiplexed over FR physical port 1 4 FR 0 X 25 UPMx XIM Four FR virtual physical ports 0 to 3 multiplexed over FR physical port 0 3 FR 1 X 25 UPM3 XIM_ Three FR virtual physical ports 0 to 2 multiplexed over FR physical port 0 plus one protocol port via physical port 3 2FR 2 X 25 UPM3 XIM_ Two FR virtual physical ports 0 to 1 multiplexed over FR physical port 0 plus two protocol ports via physical ports 2 and 3 1 FR 3 X 25 UPM3 XIM_ One FR virtual physical port 0 multiplexed over FR physical port 0 plus three protocol ports via physical ports 1 to 3 X 25 XK 75 UPM3 SA Six protocol ports via physical ports 0 to 5 2 X 25 4 FR M Two protocol ports via physical ports 0 and 1 plus UPM3 SA_ four FR virtual physical ports 2 to 5 multiplexed M over FR physical port 2 3 X 25 3 FR Three protocol ports via physical ports 0 to 2 plus three FR virtual physical ports 3 to 5 multiplexed UPM3 SA_ over FR physical port 3 4 X 25 2 FR M Four protocol ports via physical ports 0 and 1 plus two FR virtual physical ports 4 and 5 multiplexed over FR physical port 4 6 FR O X 25 UPM3 SA_ Six FR virtual physical ports 0 to 5 multiplexed over M FR physical port 0 4 FR 2 X 25 Four FR virtual physical ports 0 to 3 multiplexe
83. a CUG then the port will be treated as having outgoing access i e the call would be treated as an ordinary call 4 6 2 Access Levels within CUGs Access levels can be on a per port basis Incoming and or Outgoing Access or on a per CUG basis i e Incoming Calls Barred or Outgoing Calls Barred The Outgoing Access facility is a feature of X 25 1984 On a per port basis e Incoming Access A port can accept calls from ports belonging to other CUGs having outgoing access or from ports belonging to the open part of the network i e belonging to no CUGs at all e Outgoing Access A port can make calls to ports in other CUGs having incoming access or to ports belonging to the open part of the network e No External Access The port is not allowed to call out of its CUG X890 304751 Issue 1 4 23 Rev 0 On a per CUG basis e Incoming Calls Barred If a port in a CUG subscribes to this facility it will reject calls from other members of that CUG e Outgoing Calls Barred A port will be prevented from making calls to other members of that CUG e Neither Incoming nor Outgoing Calls Barred The port is free to communicate with other members of the CUG In Xpress terminology this is referred to as Two way Access Any combination of the per port and per CUG access levels is permitted e g Incoming Access with Outgoing Calls Barred Within a CUG different ports can have different access permissions A detailed breakdown of access pe
84. a frame relay network circuit end point Xpress uses the DLCI to assign received frame relay frames to the correct logical port and the frame relay network uses it to assign transmitted Xpress traffic to the correct frame relay circuit and hence the correct remote frame relay port The valid range for DLCls is 16 to 991 with a default of none Congestion Monitoring Period This is the period over which the ratio of frames received with and X890 304751 Issue 1 3 15 Rev 0 witnout tne Backward Explicit Congestion NotITICatION DECN DIT Set is calculated This ratio gives an indication of the business of the frame relay link and consequently the likelihood of frames being discarded by the frame relay network Xpress will automatically instigate congestion avoidance procedures whenever necessary to maintain the optimal quality of service over the frame relay network This parameter does not in any way relate to the Congestion Monitoring feature described in 3 4 6 and Appendix M 3 4 4 Network Level Parameters These parameters control the network packet level software and the establishment of switched and permanent virtual circuits SVCs and PVCs They are totally independent of whether the port being configured isa real physical port or a frame relay virtual physical port and are configured using the Configuration Port Configuration Trunk or X 25 Port Configuration Network Level Configuration screen e Logical Channel Numbers The X
85. ame relay is a streamlined protocol which effectively moves the work of handling link errors etc away from the network and gives the job to the end systems This means that the network has a much simpler job at each inter node hop as it can deal with link errors simply by throwing erroneous frames away and leaving it to the end systems to sort out the resulting protocol errors This obviously means that the network can run significantly faster with the same amount of processing power The disadvantage is that errors are more expensive to correct as they are handled by retransmitting across the entire network rather than across a single hop Hence the need for reliable links in a frame relay network As is often the case with datacomms there is frame relay and frame relay The original CCITT definition of frame relay is in ISDN terms as an Additional Packet Mode Bearer Service where user switched frame relay connections can be dialled up on the B or H channel of an ISDN link using LAP D The so far more widely adopted system is based around the original American ANSI definition of frame relay which is a PVC based system using a simplified management interface called the LMI to notify the attached DTEs of the status of frame relay PVCs pre configured by mutual agreement between the network and user This latter mechanism is the one supported by a group of organisations known as the Frame Relay Forum and is the one adopted by Series 8000 F
86. and logical port 0142 maps to physical port 2 on slot 14 X890 304751 Issue 1 0 3 Rev 0 X890 304751 Issue 1 0 4 Rev 0 Contents 1 Introduction 1 1 2 Getting Started 2 1 2 1 Powering Up the PSE 2 1 2 1 1 Booting Up the System 2 1 2 1 2 What the PSE Does to Boot Itself Up 2 1 2 1 3 If Boot up Fails 2 2 2 2 ntroduction to the Node M anager 2 3 2 2 1 Access to the Node Manager 2 3 2 2 2 UsingtheMini PAD 2 4 2 2 3 Logging on tothe Node Manager 2 5 2 2 4 Menus and Commands 2 5 2 2 5 Useof the Function Keys 2 6 2 2 6 OnlineHelp 2 6 2 2 7 Logging Out 2 7 2 2 8 Special Characters 2 7 2 3 Management of Applications 2 8 2 3 1 Configuration of Applications Management 2 8 2 3 2 Access to the Management Screens of an 2 9 Application 2 3 3 Logging Out from an Application 2 9 3 Configuring a Single PSE 3 1 3 1 Physical and Logical Links and Ports 3 1 3 1 1 Physical Links and Ports 3 1 3 1 2 Logical Ports 3 2 3 1 3 Frame Relay Physical and Virtual Physical Ports 3 3 3 2 Logical Port Allocation 3 5 3 3 Module Configuration 3 6 3 3 1 Module Parameters 3 6 3 3 2 Display Version Numbers 3 9 3 3 3 Change Module Link States 3 9 3 3 4 Module Restarts 3 9 3 4 X 25 X 75 F rame Relay Port Configuration 3 10 X890 304751 Issue 1 0 5 Rev 0 34 1 FNYSICal Level Farameters 3 4 2 Frame Relay Core Level Parameters 3 4 3 Data Link Level Parameters 3 4 4 Network Level Parameters 3 4 5 User Facilities 3 4 6 Congestion Monitoring 3 4 7 Error Monitorin
87. apped from X 25 User Facilities a Throughput Class Indication This is the same as X 25 Throughput Class Negotiation except that the lower of the default and requested values is assumed if the called network does not respond b Packet Window Size Indication This is the same as X 25 Flow Control Parameter Negotiation except that the default not the requested values are assumed if the called network does not respond c Fast Select Indication This is the same as the X 25 facility d Reverse Charging Indication This is the same as the X 25 facility e CLAMN This is the same as the X 25 facility Xpress transfers the CLAMN utility transparently over X 75 1984 interfaces and maps it back into a User Facility at an X 75 1980 interface Notes Xpress routes on RPOA selections carried in the X 25 User Facilities field of X 75 call request packets Xpress also provides RPOA Subscription at X 75 ports Xpress does not support international CU Gs but does pass them transparently when it acts as a transit network In all other cases Xpress rejects any call request at an X 75 link which holds an X 25 or X 75 CUG selection except when it acts as an originating network and there is a CUG with Outgoing Access selection In this case Xpress removes the selection and forwards the call request packet Other Differences Between X 75 1984 and X 25 1984 88 At X 75 interfaces Xpress always clears back the outgoing call when
88. arity DLCI etc Xpress Trunk Protocol Data Unit The Xpress Protocol Data Unit is encapsulated in the frame relay frame Figure A 9 Trunk Protocol Encapsulation X890 304751 Issue 1 A 27 X890 304751 Issue 1 A 28 Rev 0 Appendix B Error Causes and Diagnostic Codes This appendix lists the clearing causes resetting causes and restarting causes issued by Xpress It also lists the diagnostics codes B 1 Clearing Causes X890 304751 Issue 1 Clearing Cause DTE Originated Number Busy Invalid Facility Request Network Congestion Out of Order Access Barred Not Obtainable Remote Procedure Error Local Procedure Error RPOA Out of Order Reverse Charging Acceptance Not Subscribed Incompatible Destination Fast Select Acceptance Not Subscribed Table B 1 Clearing Cause Codes B 1 Rev 0 B 2 Resetting Causes Resetting Cause DTE Originated Out of Order Note 1 Remote Procedure Error Local Procedure Error Network Congestion Remote DTE Operational Note 1 Network Operational Note 2 Incompatible Destination Network Out of Order Note 1 Table B 2 Resetting Cause Codes Note1l These resetting causes are reserved for PVCs Note2 This resetting cause indicates that the PSE has been able tore establish an SVC or PVC X890 304751 Issue 1 B 2 Rev 0 B 3 Restarting Causes Restarting Cause DTE Originated Local Procedure Error Network Congestion Network Operational
89. ary choice This forces them to re establish with the result that these previously displaced calls can be periodically returned to their primary route trunk port if it is available This may be very useful if Automatic Port Reinstatement has been configured to take place on the primary Automatic Port Reinstatement is enabled by entering a non zero value for the Port Reinstatement delay parameter located on Configuration Port Configuration Trunk Port Configuration Error Monitoring The values of the two parameters should be considered jointly it is wasteful for calls to try frequently to re establish on a primary while it remains closed for a long period which would be the case if the Port Reinstatement delay is set many times longer than the Auto Reroute Interval set on the secondary Conversely the opportunity to return calls to the primary as soon as it is reinstated is lost if the Auto Reroute Interval is too long Notes Trunks always support a maximum data packet size of 512 octets All User Facilities are transferred transparently across trunks A trunk port s packet window sizes are not affected by flow control parameter negotiation carried out at X 25 ports Trunk ports always support the Extended F ormat of packets All Fast Select and Reverse Charging requests are forwarded across trunks Trunk ports do not affect the support of D bits Auto reroute should be symmetrically configured at the two ends of a trunk
90. ash twice then the previous dump file for that module would be overwritten so any dump files should be archived immediately onto a spare disk using the File Copy screen and returned to your supplier as soon as possible If the system fails to dump a module when it crashes 1 Check that the auto dump flag is correctly set If not enable this flag by using theConfiguration Module Configuration Edit Module Parameters screen 2 Check that you have an appropriate dumper program dmp Upm Um L dmp U03 Um L or dmpU03Xrmc L on your disk 5 4 1 Delete Dump File This screen enables you to delete a dump file from the dump disk Dump files are normally stored on the disk in drive b To check the contents of the disk use the Utilities Disk utilities List File Directory screen Section 5 3 3 Please check that a dump file is no longer needed before it is removed deleted files cannot be recovered 5 4 2 Print Dump File This screen lets you print the contents of a dump file The dump file to be printed is identified by its bay and slot number Dump files are normally stored on the disk in drive b the right hand drive Dump files are quite large and will consume a lot of paper so it is recommended that you don t print them unnecessarily X890 304751 Issue 1 5 10 Rev 0 5 5 Install Delete E xpand Applications This screen allows you to administer the applications which are installed ona PSE 5 5 1 Display When you first enter the
91. ation Name This option allows you to select which application you wish toload and run on a module When you select the option the Node Manager will display a list of the available applications which may be run on that type of module There are two basic types of applications Native and Imported Native Applications This type of application is always available as they are distributed as part of the standard software release which is bundled with every Xpress node X890 304751 Issue 1 3 6 Rev 0 I Ne currently avalladie native applications inciuae tne Node Manager the X 25 X 75 Trunk protocol suite and various X 25 X 75 Trunk over frame relay combinations Which of the latter is selected defines which ports on a card run X 25 X 75 Trunk directly over their physical interface and which run the requisite protocol over a frame relay virtual physical link as described in Section 3 1 3 A list of selectable native applications together with their requisite card type s is given in Tables 3 1 and 3 2 Note that further mixes of ports may be available and will follow the general layout given in the tables For details of the card types refer to the appropriate installation guide In the tables the term protocol port refers to an X 25 X 75 Xpress inter node trunk port APPLICATION CARD S DESCRIPTION Node Manager The node manager software Crash Dump Dumps the node manager should it fail Do NOT select this application manually X 25 K 75
92. ay can be disconnected from the hardware and multiplexed over port 0 to the frame relay network for connection to six different destinations via remote frame relay network ports In this case the logical port numbers assigned to physical ports 0 2 0 to 0 2 5 are referencing frame relay virtual physical ports In addition the logical port number assigned to physical port 0 2 0 is also referencing the real frame relay physical port 0 2 0 This mechanism is totally independent of logical port numbering In the example given above it is the physical ports 0 2 0 to 0 2 5 which are multiplexed over port 0 2 0 and not the logical ports 0020 to 0025 assuming direct logical to physical mapping has been used This means that if logical port 0020 were moved to a physical port on another card then the virtual physical port mapping would not follow that logical port but would remain associated with physical port 0 2 0 and hence with whatever logical port is assigned toit It is in fact not possible to move a logical port which references an FR virtual physical port to a real physical port and vice versa The management software polices the Suitability of all logical to physical and logical to FR virtual physical port assignments This mechanism is equally applicable to Xpress trunks See Chapter 4 which can be freely mixed with ports and carried over frame relay in the same manner X890 304751 Issue 1 3 4 Rev 0 3 2 Logical Port Allocation Wh
93. ber of the node on which the port to which this address refers resides If the node number field is not equal tothe node number of this node then the rest of the address is ignored and the node number field is used to derive a route to the required node as explained in Chapter 4 Digits 8 to 11 are the logical port number of the required port X890 304751 Issue 1 3 27 Rev 0 Digits 12 to 14 are the sub address field This field is optional and can be 0 1 2 or 3 digits long The sub address plays no part in call routing within Xpress PSEs and is simply passed transparently to the attached device where it may be used to identify for example a port on a PAD or an application on a host For example the address used to make a call to port 2 on an async PAD connected to logical port 0160 on node 1 within this Xpress network could be any one of 1100 00101602 1100001016002 1100001 0160 002 The DNIC is the Xpress escape DNIC the node number is 001 the logical port number is 0160 and the sub address indicating async port 2 can be 2 02 or 002 depending on the format required by the PAD Xpress always insists on at least the first 11 digits of this scheme being present in the correct format and order so as to be able to internally route a call However this does not mean that externally attached devices must also stick exclusively to this format Xpress provides powerful address translation and analysis functions which allow virtually any ad
94. bits the PSE will reroute calls transparently without any loss of data In addition it is possible to configure a trunk to clear all calls across it that are not using the trunk as their optimum choice In this case these calls will be automatically re established as described above taking the current best route available This is useful especially in the case of slow and or expensive backup trunks such as dial up trunks to ensure that calls are rerouted using the optimal route should a trunk fail and then subsequently become available again See Section 4 3 1 4 For PVCs the PSE signals that a trunk has failed and it is attempting to reroute the VC by generating resets with the Cause Code Network Out of Order at the X 25 ports For both SVCs and PVCs the PSE signals that the VCs have been successfully rerouted by generating resets with the Cause Code of Network Operational at the X 25 ports See Appendix B for information about Cause Codes X890 304751 Issue 1 4 2 Rev 0 4 2 Node Numbering The node number and node location name are configured by accessing the Configuration Node Configuration Edit Node Identity menu Each PSE is assigned a node number and a node location name The node number must be in the range 0 999 The default valueis 1 The node location name may be up to 20 characters long It is for user information only and is not used by the PSE The node number and node location name both appear on the top lin
95. by the connected device If ICIS set to internal NE FSE WII supply the line clock using its internal Baud Rate Generator The default is External e Clocking Speed This can be set to a range of values between 2400 and 256000 bps Note that it must always be set even if the PSE is using the externally supplied clock This is to allow the software to give sensible values to some of its internal timers and to enable it to calculate link utilisation correctly on which Congestion Monitoring depends It should be set to the speed of the connected device if known or 2400 bps otherwise The default is dependent on the ACM Type The aggregate clocking speed for a module is dependent on the UPM type 64000 bps UPM1 and UPM2 256000 UPM3 UPM4 and SAC e Physical Interface This parameter must be set to reflect the type of Physical Port to which the Logical Port is assigned This item is completely hardware dependent so if a Logical Port is ever moved between Physical Ports on different types of XIM this parameter must be updated to ensure correct operation Choices are V 24 XIM1 X 21 XIM2 V 35 XIM3 V 36 XIM2 V 54 XIM1 e Transmit Flag Insertion The PSE s link protocol software is capable of transmitting information frames with a single inter frame separator or flag for sustained high speed operation This can cause problems with some non Cray equipment which cannot receive frames this fast If this parameter is set to YES the s
96. call collision occurs X890 304751 Issue 1 A 19 Rev 0 X 75 does not support diagnostic or REJ packets Clear request packets always contain Address Network Utility and User Facility fields Clear confirmation packets always have the basic format i e the Packet Type Identifier byte is the last byte of the packet CUG amp CUG OA selections are used in conjunction with international interlock codes X 75 does not support data packet sizes larger than 1024 bytes Instead of DTE and DCE timeout values T20 23 amp T10 13 respectively both X 75 STEs have the same timeout values T 30 33 These timeout have the same values as T 20 23 If Xpress issues a clear reset to an X 75 interface then the cause code is usually set to network congestion see Appendix C of X 75 1984 If a clear reset does not originate at the local X 75 interface and the cause code indicates network congestion then Xpress ensures that the diagnostic code indicates either no additional information or international problem Otherwise Xpress transparently transfers diagnostic codes across and X 75 interface A 2 7 3 Differences Between X 75 1980 and X 75 1984 This section gives an overview of the differences between X 75 1980 and X 75 1984 at the Packet Layer The 1980 version restricts interrupt request packets to one byte of User Data whereas the 1984 version allows up to 32 bytes The 1980 version restricts the si
97. call is quite likely to throw the Call Request packet away and leave the caller to time out and clear down 3 7 6 Failed Installation of an Application There are two ways that an installation can fail assuming that the system disk and distribution disks contain all the necessary files and are not corrupt X890 304751 Issue 1 3 40 Rev 0 1 theapplication is already installed onto the PSE In this case you must delete the application from the PSE before attempting to install a new version of the application See Section 5 5 3 for details of how to delete an application 2 the system disk becomes full during the installation In this case the Node Manager will cleanly abort the installation You must then free up space on the system disk before making another attempt to install the application X890 304751 Issue 1 3 41 Rev 0 3 8 Hunt Groups Hunt groups are a feature provided by CCITT Recommendation X 25 1984 Logical ports on a PSE may be clustered into hunt groups The hunt group mechanism allows individual X 25 calls to contend for connection to one of a number of X 25 ports If a call specifies a hunt group address then it is forwarded to the next available port in the group Hunt groups thus provide primitive call balancing spreading the load over the member ports Members of a hunt group must be ports on a single PSE node thus a hunt group cannot span a number of nodes Each node is limited to a total of 16 hunt gr
98. ce of characters and hence files Pattern matching is not allowed if the from file is on a remote node The pattern matching characters are X890 304751 Issue 1 5 7 Rev 0 r matches any sequence of characters matches any single character SET matches any single character in the specified SET SET matches any single character not in the specified SET A SET consists of single characters or a range A range is two characters separated by a hyphen e g 0 9 to match any single digit 0 to 9 A Z to match any single character A to Z Some example file names are a help data a applic data a nmU03Um L a x25 config a123 help data and some example patterns to match the above files are a a to match all files ending in an a i e the data files a aL tomatch all data and L files i e all the data and load files a g tomatch all files that do not end in the character g Some example commands for local node file copies would be Copy a config b this will copy all the system configuration files to drive b This is useful for backing up your configuration and having the files online Copy a g b this will copy all files other than the configuration files to drive b Copy b core 0 1 a save dump this will copy the dump file from drive b to drive a changing the name in the process Some example commands for remote node file copies would be Copy a x25UpmXim L a123 this will copy the X 25 application on the l
99. ch module raised it and a brief summary of the problem Full details of the problem as logged to the printer are available by using the Expand command Alarm or Warning events may need operator intervention Therefore the status of an event from the time it occurs until it is fixed is displayed and should be maintained by the operator using the commands available When an alarm or warning occurs it is given the status NEW When you ACKNOWLEDGE it its status changes to CURRENT When the cause of the event is corrected you can CLEAR it and its status changes to CLEARED You can subsequently delete all on line record of that event by using the DELETE command Sometimes the system will detect that the problem has cleared itself e g a link going down then coming up again In this case its status will automatically be moved to CLEARED once you ACKNOWLEDGE it X890 304751 Issue 1 5 16 Rev 0 5 8 Charging Charging is a feature of X 25 1984 and X 25 1988 It can be configured per port as part of the Configuration Port configuration X25 port configuration User facilities screen Every call made by that port will cause charging information to be sent to it in the Clear packet The charging information record contains a subset of the information in a billing record as shown in below call duration 1 4 Binary Call duration in four 8 bit fields format DD HH MM SS Number of segments 1 number of segments received by DTE segment 6
100. cleared down once the last client disconnects The ABS server is dormant when there are no clients connected This saves network charges for example where the host is accessed via a public network or is provided by an external agency When the first client connects to a server that server makes an X 25 call to a fixed address The address is of the format 1100 nnn 9ss9 995 where nnn the node number and ss the slot number of the card on which the server resides In order to map this address to the required host address the Address Analysis mechanism described in Section 4 5 1 is used For example if the server is in node 140 slot 7 and the host is connected to asynchronous port 3 of a PAD connected to logical port 45 on the same node then an entry should be made in node 140 s AAT as follows Match Address Internal Address 1100 140 9079 995 1100 140 0045 003 X890 304751 Issue 1 G 4 Rev 0 This mapping will cause the node to internally replace the fixed address used by the server with the address of the host and consequently route the host call to the correct PAD port Note that the called address field in the call request packet generated by the server is not changed by the address analysis procedure and that consequently the PAD will receive the original server called address of 1100 140 9079 995 It is necessary either to use outgoing called address translation Section 4 5 2 on port 45 to translate the called address as required
101. cribes how to configure and use an Xpress PSE The facilities and features described are those offered by Version 9 of the PSE software Chapters 1 and 2 provide an overall view of the PSE and should be read thoroughly before attempting to use it Chapters 3 4 and 5 describe in more detail the configuration and management of the PSE Chapter 6 contains useful information on what can be done in the event of problems The appendices provide a description of the X 25 frame relay protocols a glossary ACS support and other information that may be needed for reference Conventions Used in this Guide The PSE node Manager is a menu driven system and throughout this guide there are dialogues between the user and the PSE menu system These are represented by different fonts as follows Terminal displays and printouts are represented as e g Terminal Displays and Printouts User responses are represented as e g User Responses Single keystrokes are represented as e g PF1 Double keystrokes are represented as e g CTRL T Non literals are represented as e g lt timeand date gt The route by which a particular action screen is reached is represented starting from the main menu as e g Configuration Port Configuration X 25 Port Configuration Throughout this guide logical port numbers are assumed to map directly to physical port numbers according to the common convention For example logical port 0030 maps to physical port 0 on slot 3
102. ctly if DCD is held either permanently high or permanently low as long as RI is used to indicate an incoming call if required RI and DCD can be used together as long as RI going low precedes DCD going high for an incoming call K 2 3 V 11 Interface Circuits The signalling used to drive a V 11 dial up interface is based on X 21 control signals as follows C Pins 3 and 10 Control C is raised by the node to signal that the attached DCE should dial a pre defined number to establish a connection and is dropped by the node to signal that the connection should be dropped X890 304751 Issue 1 K 4 Rev 0 Note that when the connection is not required the node will signal Control off and transmit a continuous stream of ones i e it will be signalling Ready When the circuit is required Control will be raised and HDLC flags transmitted i e the interface will switch from Ready to Data Transfer In order to clear the circuit the node will drop Control and revert to transmitting all ones thus switching from Data Transfer back to Ready I Pins 5 and 12 Indicate is raised by the DCE toindicate An incoming call Successful establishment of an outgoing call i e raised in response to Control after connection sequence is completed Dropped by the DCE to indicate circuit failure or call cleared by the remote end Note that the dial up software does not rely on Indicate going high to detect a link establishment this it does by detec
103. cts billing information at regular intervals from each of the XI Ms in turn so there is a time delay between the call being cleared and billing information being forwarded to the billing destination Billing record collection can be turned off and on again under operator control The connection to the billing destination is itself an X 25 SVC Ifthe connection to the billing destination cannot be brought up or is lost the Node Manager will try to re establish the call at regular intervals If it cannot be re established billing data will start to back up When the backlog reaches a predefined limit billing records will start being discarded to conserve memory An alarm is generated if this happens X890 304751 Issue 1 5 18 Rev 0 As stated earlier two billing records are generated for each completed call one from each end of the call Even if calls are made from the Xpress PSE to an external network two billing records will still be generated one from the calling Xpress PSE port and the other from the gateway port where the call actually entered the external network PVC calls do not normally cause billing records to be generated the exception is when a port at which the PVC channel is configured is put out of service 5 9 2 Configuration There are two parts to configuring a billing destination 1 You must specify its X 121 address 2 You must specify the conditions under which you require billing collection to take place On
104. d over FR physical port 0 plus two protocol ports via UPM3 SA_ physical ports 0 and 1 3 FR 3 X 25 M Three FR virtual physical ports 0 to 2 multiplexed over FR physical port 0 plus three protocol ports via UPM3 SA_ physical ports 3 to 5 2 FR 4 X 25 M Two FR virtual physical ports 0 and 1 multiplexed over FR physical port 0 plus four protocol ports via physical ports 2 to 5 UPM3 SA M X890 304751 Issu 1UPM3 SA Table 3 2 8425 8525 Native Applications mported Applications Imported applications are distributed separately to the core Xpress operating software and must be explicitly installed from a distribution disk set Details of the imported application installation procedure are given in Chapter 5 The Node Manager will automatically restart the module when PF1 is pressed to submit the edit module configuration screen once the application has been selected It will then load the module with the selected application Note that the node manager will not allow the edit module configuration screen to be submitted with PF1 if there are any ports on that module that are not out of service The selected application will be recorded on disk so that the Node M anager can automatically load the correct application whenever the module or node restarts 3 3 2 Display Version Numbers This option allows the issue and revision numbers of the UPM and ACM cards in a specified bay and slot to be displayed together with the version
105. d congestion monitoring period parameters of any frame relay virtual physical ports iv Set up the Network Layer of all X 25 X 75 and Application ports v Set up the User Facilities of all X 25 X 75 and Application ports 5 Bring the ports into service by setting the port state to on line This can be achieved using one of the following menus Configuration Node Configuration Change state of all ports on node Configuration Module Configuration Change module link states Configuration Port Configuration X 25 X 75 Application port configuration Change State of port Configuration Logical Port Configuration Change state of a logical port 6 At some later time when basic operation of the node has been achieved inspect the Utilisation level and mean error rate on the ports by means of Configuration Node Configuration Detailed Link Status Display For any ports on which Congestion or Error Monitoring are required to operate put the ports out of service and configure the features as appropriate Notes a Ports are always in out of service state when they are first created A port must bein this state in order to have its configuration edited In addition to edit the Physical or Frame Relay Core level parameters of a frame relay physical port all of the frame relay virtual physical ports multiplexed over it must also be out of service b In addition to on line or out of service a port may also be set to off
106. d for use after a configurable period of time has elapsed An event is generated whenever a port closes It contains the error rate at the time of closure and indicates whether or not the port is going to be automatically re instated later Another event is raised whenever automatic reinstatement takes place It is recommended that Error Monitoring and Control is not applied toa port unless it has an alternative e g it belongs to a hunt group or is a trunk with a secondary etc Otherwise whenever automatic closure of the port occurs the affected network users will suffer complete loss of service The mean error rate value that is used by this feature is available for inspection on the following Node Manager status screen Configuration Node Configuration Detailed Link Status Display screen The calculation depends on a minimum amount of line activity and so the value displayed shows as 1 or less if a port is out of service closed due to errors or down On this and other status screens the port state errs is shown for a port temporarily closed due to errors distinguishing it from ports that are down for other reasons The parameters on the screen are described below e Error tolerance limit This defines the maximum mean error rate that the system will tolerate on this port Whenever the measured error rate exceeds this limit the port automatically closes itself The default value is 100 Experience shows that a trunk or lin
107. d format of RPOA Selection The Extended F ormat may hold up to three RPOA Selections See Section 3 4 4 20 21 NUI Subscription Xpress provides this 1988 facility by means of two configuration options Local NUI Selection and Remote NUI Selection These two options allow the operator to specify one of the following at a port transfer any NUI selection transparently when present reject all requests which don t hold an NUI reject all requests which do hold an NUI X890 304751 Issue 1 A 14 Rev 0 A 2 6 Calls Between X 25 1980 and X 25 1984 1988 Ports Xpress allows calls between X 25 1980 and X 25 1984 1988 ports To avoid violation of X 25 1980 during such a call Xpress will Clear the call if the size of a packet s User Facilities field exceeds the maximum allowed by X 25 1980 Clear the call if a packet holds an extended CUG selection transit delay selection and indication facility CCITT specified DTE facility or any other 1984 1988 specific User Facility Forceall cause codes to be consistent with X 25 1980 before sending them to a 1980 port Reset the call if an Interrupt packet holds a user data field larger than that allowed by X 25 1980 A 2 7 Xpress and the X 75 Recommendation This Appendix describes how Xpress supports X 75 The description is mainly in terms of the differences between X 75 and X 25 A 2 7 1 Introduction The CCITT publishes the X 75 Recommendation
108. d locally or by an NMC connected to the network The required event handling and state of the node can be selected on the Configuration Node configuration Edit node configuration screen When the node state is changed from on line to off line the events will no longer be forwarded tothe NMC This enables maintenance to be carried out on a node without flooding the NMC with spurious events All events are sent to the printer to give a permanent log and are labelled with their time of occurrence Serious events are preceded by their severity There are three classes of events of increasing severity e Ordinary These events are generally of an administrative nature and enable the operator to maintain a log of normal changes to the system such as an operator logging in changing a configuration and logging out Ordinary events do not result in the operator having to take some action e Warnings These normally indicate system degradation which may be caused by an incorrect configuration or exceptionally high load eg a module running out of memory for buffering calls e Alarms These normally indicate failure of a component or a resource e g a module or link failure or the printer being off line An alarm will usually require immediate operator intervention to correct Warnings and Alarms represent exceptional conditions within the system and should prompt the operator into remedial action It is not possible to disable events and
109. dit the application name 5 Typethe correct number to select the FR 012345 application six Frame Relay virtual physical ports multiplexed over port 0 6 Type PF1 to submit the screen The selection of the Frame Relay application will be confirmed and slot 7 will be re started 7 Type PF4 to return to the main menu 8 Select Configuration Node configuration Node status display 9 Notethat the Operational State of slot 7 is Loading Updatethe screen by typing r RETURN until the operational state changes to Operational 10 Type PF4 to return to the main menu Step 2 Assign the logical ports 11 Select Configuration Logical port allocation Create a new logical port 12 Type t70 RETURN to configure logical port T0070 13 Type ph RETURN 7 RETURN 0 RETURN to set the virtual physical port number to bay 0 slot 7 port 0 X890 304751 Issue 1 4 11 Rev 0 14 Type po RETURN FR trunk to node 2 RETURN to set the port description Submit the form with PF 1 Typer RETURN t71 RETURN to create logical port T0071 Repeat steps 13 to 15 but with a port number of 1 and FR trunk to node 3 as a description 15 16 17 18 Hit PF3 twice to go back to the Configuration menu Step 3 Configure the Frame Relay physical interface s physical and core data link layers This is done via logical port T 70 19 Select Port configuration Trunk port configuration Physical l
110. dress format to be used by attached devices the Xpress node mapping the external addressing scheme to the internal one automatically These mechanisms are described in Section 4 5 3 5 2 SVC Configuration As explained above SVCs require no additional configuration once the ports to which external devices have been given logical port numbers and have been suitably configured The internal addressing scheme and address translation facilities make explicit configuration of device addresses unnecessary 3 5 3 PVC Configuration 1 Define on each port the range of logical channels used for PVC calls see Section 3 4 3 2 For each PVC on a port identify the remote end of the call by specifying the remote node number logical port number and logical channel number This configuration process must also be carried out at the remote PVC so that each end of the PVC knows the identity of the other For example X890 304751 Issue 1 3 28 Rev 0 PVC on port 0020 on channel number 2 End to end delivery confirmation no Remote end of PVC Node 3 Logical port 0030 Channel number 2 from DTE to DTE Window size 7 7 Max packet size 128 128 Each PVC call may be configured with its own values for maximum packet size and window size By default the PVC will inherit the values for these parameters from its logical port configuration 3 5 4 PVC Call Establishment The virtual circuit completing the PVC is established automatically by th
111. dress will have been overwritten Note Call Request Containing RPOA will also have Address Translation done Exit Port X890 304751 Issue 1 4 15 Rev 0 4 4 3 The Routing Table The routing table tells the PSE how to route calls to other nodes in the same Xpress network A single node network doesn t need one The routing table is indexed by the destination node number It lists up to three trunk ports over which a call should be forwarded to reach the destination node there is no entry in the routing table for the local node For example Routing Table Entry for destination node 1 Primary exit port T0090 Secondary exit port T0050 Tertiary exit port T0020 Maximum projected Hop count 2 For a call to traverse a network of PSEs successfully a correctly configured routing table must exist at every node en route to the destination node Wildcards can be used in the routing table to reduce the number of entries required The node number range is 000 999 wildcards are specified by replacing one or more of the digits with an n For example if nodes 100 101 102 and 103 are all best reached by the same set of next hop trunks but node 104 is reached via a different route then an explicit entry of 104 will handle the calls for that node and a wildcard entry of 10n will handle calls to the other nodes Note that any wildcard entries always follow any explicit entries in the routing table If the node only has one route
112. dware when the PSE is powered up or when a logical is moved to another physical port For example logical port 1234 may initially refer to physical port 0 2 1 but can be easily and quickly re assigned to another physical port if necessary All the configuration apart from the physical port characteristics of port 1234 is automatically assigned to the new physical port when the move takes place The assignment of particular logical port numbers to physical ports is largely the user s choice As an example it may suit a particular installation to reflect the physical port number in the logical port number e g Physical Port Number Logical Port Number Bay 0 Slot 1 Link O 0010 Bay O0 Slot 1 Link O 0011 Bay 0 Slot 4 Link 3 0043 Bay 0 Slot 12 Link 5 0125 This convention has been adopted by the vast majority of network administrators but nevertheless it may make sense to number logical ports according to some departmental or company wide numbering scheme However the restrictions within the 0000 9999 range given in Section 3 2 should be noted X890 304751 Issue 1 3 2 Rev 0 3 1 3 Frame Relay Physical and Virtual Physical Ports One of the possible alternatives to a simple direct connection between an Xpress physical port and an attached X 25 device is a connection via a frame relay network This mechanism is mentioned here as it involves the multiplexing of the traffic from multiple frame relay virtual physical ports over a si
113. e PSE when the X 25 ports are brought on line In a busy network this may not happen instantaneously but may take a few seconds to complete If the PVC connects ports on different nodes then the routing table must be correctly configured in order for the connection to be successfully established X890 304751 Issue 1 3 29 Rev 0 3 6 Examples of Port Configuration This section is intended to take a first time user step by step through the procedure required to configure logical ports on a PSE node such that a call may be made between the devices attached to the ports All simple non trunk port configurations will follow this procedure closely so that once the procedure is understood no problems should be encountered in configuring a full node An example of frame relay port configuration is given in Section 4 3 2 3 6 1 Example 1 X 25 Port Configuration This first example assumes that two PADs are to be connected to physical ports 0 1 0 and 0 4 2 on XIM1s V 24 Remember that you are not configuring the PADs but the ports to which they are connected This means for example that if a PAD is configured as a logical DTE then the port must be configured as a logical DCE This example assumes that no logical ports have been set up on the node The PADs have the following configuration PAD 1 0 1 0 PAD 2 0 4 2 Electrical Interface DTE DCE Line Speed External 19 200 bps Level 2 Window Size 7 7 Data Link Interface DTE DTE
114. e allowed to do this use the Access utilities User access specification Edit user screen 5 1 2 Type Specification Eight user types exist numbered from 1 to 8 Type 1 is the highest category and is reserved and unchangeable It is allocated to the user wizard A profile consists of a set of access rights to each system function An access right may be one of the following No access Read Only access Update access The services to which you may have access rights are Alarms Warnings Statistics User access specification System utilities Logical configuration Physical configuration Billing specification Routing specification Management of applications Once the profile has been constructed it can then be applied to any number of new or existing users The profile is identified by the number assigned to it X890 304751 Issue 1 5 2 Rev 0 5 1 3 User Access Menu This menu lets you select the user access specification function that you need These functions allow the manager to add to delete edit and display the list of valid user logon names 5 1 3 1 Create User This screen enables you to enter a new user into the database User Name and Password are those used by the user to log on The Name and Password can be up to 16 characters long For security purposes the Password is always displayed as a series of asterisks The User Type must bein the range 1 to 8 This number refers to a profile previously set
115. e application Goto the create a new logical port screen and type in the logical port number for the application Use logical port 7003 Type RETURN at the end Type ph RETURN Type 3 RETURN 0 RETURN to select slot 3 link O X890 304751 Issue 1 3 35 Rev 0 10 11 12 13 Type po RETURN My application RETURN to set the port description Press PF1 to submit the screen The message Logical port created successfully will be displayed If it isn t refer to Section 3 7 3 below To set the X 25 configuration of the application link press PF3 to return tothe Logical Port Allocation menu Enter the logical physical port display screen to check that all is well This screen displays the allocated logical port numbers together with the Bay Slot and Link numbers to which they are assigned and their port descriptions Press PF3 twice to return tothe Configuration menu and enter the Port Configuration X 25 Port Configuration menu Application links do not have physical or data link level configuration Enter the network level screen and type in the application link s logical port number 7003 Type t RETURN to set up the Two Way LCN range Type I RETURN 1024 RETURN to set up the lower LCN Type u RETURN 1087 RETURN to set up the upper LCN Press PF2 to escape from the boundary change prompt Therest of the configuration is correct so press PF1 to submit the screen The message Co
116. e of the display The default node location name is Watford X890 304751 Issue 1 4 3 Rev 0 4 3 Trunks A trunk is an inter node link which is attached toa port on a XIM There can be many trunks between two Xpress PSEs Xpress PSEs are linked together with trunks to form networks A trunk port is physically the same as an X 25 port but is configured in a slightly different way Trunks operate in the same way as X 25 1984 at the Physical and Data Link Levels but in a different way at packet level The Packet Level protocol is described in Appendix A Trunks can be carried over frame relay networks in the same way as ports as described in Chapter 3 The Xpress trunk protocol is enhanced at each version to support new features In order to permit node by node upgrades of previous software versions the trunks may be configured to operate in Backward Compatibility Mode This enables calls to be made across mixed networks but it should be noted that trunks running in compatibility mode cannot carry calls employing version specific facilities It is also possible that other network functions such as call re establishment may not operate reliably over such trunks For other reasons it is not recommended that mixed networks are used operationally other than during the transition period whilst previous version nodes are upgraded to the current software version 4 3 1 Trunk Port Configuration The procedure for configuring trunk ports is
117. e originating and destination networks are indicated by the Calling and Called addresses Xpress rejects any call set up packet in which the TNICs indicate looping X 75 does not limit the number of TNICs held in a call set up packet provided they fit within the Network Utilities field This means that a maximum of 19 TNICs may be held this number allows for the mandatory Call Identifier utility See b below X890 304751 Issue 1 A 17 Rev 0 Xpress adds its TNIC the Internetworking DNIC see Section 3 9 1 to a call request packet immediately before forwarding the packet over an X 75 interface Xpress clears the call if there is no room in the packet Xpress also provides a configurable per port option which allows the operator to suppress TNIC insertion Xpress ensures that call accept packets contain a TNIC which specifies its Internetworking DNIC b Call Identifier This is established by the originating network and when used in conjunction with the calling X 121 address uniquely identifies an SVC Transit networks transfer the call identifier without changing it When Xpress is the originating network it will assign the Call Identifier to be the same value as the SVC count which is included in Xpress Billing Records See Appendix C The Call Identifier cannot be used for identifying PVCs However a PVC can be uniquely identified by the X 121 addresses and the LCIs at the two ports c Clearing Network Identificatio
118. e provides debugging aids It allows applications to store interesting data areas if thereis a core dump The X890 304751 Issue 1 J 7 Rev 0 Apress Kernel sortware will 109 and raise an Event message whenever an application incorrectly invokes a system call Network Management The Xpress Kernel software provides a system call nmctl which allows applications to exchange messages with the NMC Applications may communicate directly with the NMC without involving the host Node Manager except for Event messages which are forwarded via the Node Manager X890 304751 Issue 1 J 8 Rev 0 Appendix K Dial up Ports K 1 Overview Previous to software Version 7 2 every X 25 X 75 link or inter node trunk was assumed to be permanently available and up at levels 2 and 3 i e the links were always assumed to be provided by digital leased circuits auto restoral synchronous leased line modems etc Software Version 7 2 onwards no longer has this restriction It is now possible to configure a link or trunk to be dial up i e the link is provided by a mechanism which means that it is not physically established until it is required to carry packet traffic Examples are links provided by synchronous V 32 dial up modems or via ISDN Terminal Adapters TAs Figure K 1 shows two possible configurations where a remote X 25 card equipped PC is connected into the network via a modem link and a dedicated inter node trunk is backed up by an ISDN link PC
119. e upgrading a PSE to V7 or later software you must first ensure that there are no physical ports with Logical port numbers in the range 7000 to 7999 X890 304751 Issue 1 3 5 Rev 0 3 3 Module Configuration This option allows you to view and configure the operating parameters for a physical module The parameters are divided into four sections e Edit module parameters e Display software and hardware version numbers e Change module link port states e Module restarts 3 3 1 Module Parameters e Dump after failure This flag controls the dumping of a failed module see Section 5 4 e Auto Load This flag controls the loading of the module If set the Node Manager will automatically load a module when it is inserted or after restarting due to a failure e Module Buffer Sizes This option allows you to view the capacity of the various memory buffers contained in the specified module The danger levels and recovery levels are preset and should not need to be changed although this can be done if necessary The size of the buffer contents varies according to the amount of traffic being handled by the module s links If any buffer does become full this information will be displayed on the Warnings screen but the full state is only temporary and will clear when the calls have been completed All ports on the module will refuse to accept further calls until the amount of data in the buffer has fallen to an acceptable level e Applic
120. ed Set thresholds and other Congestion Monitoring and Control parameters on the appropriate trunk ports Refuse thresholds are more straightforward than clear thresholds and do not cause extra events to be generated However refuse thresholds cannot control the amount of bandwidth used by a call once it has established In contrast a port with clear thresholds set is able to exercise such control by bumping calls off the trunk If Clear Thresholds are to be used for a given priority class it is often wise to set refuse thresholds for the lower priority classes as in the example Arrange for attached devices to request the appropriate Throughput Classes so that Xpress can distinguish between priorities of call Alternatively use Default Throughput class on a user port if all calls entering it are to be given the same priority Select correct Priority Class Profile at the user ports gateway ports and at any trunk ports that attach to pre Version 9 nodes Configure Auto Reroute Interval on those trunk ports that comprise the secondary and tertiary routes if periodic automatic call restoral to the primary is required Enable Network Data Integrity parameter at user ports if required X890 304751 Issue 1 M 14 Rev 0 Appendix N Error Monitoring and Control N 1 Introduction The Error Monitoring and Control features can be used in a similar manner to Congestion Monitoring and Control Allowing error rates on individual po
121. ed in conjunction with a dial up trunk as calls will be automatically removed from the dial up trunk which will then shut down and re establish down a less expensive X890 304751 Issue 1 4 6 Rev 0 route once avaliable sIMIlarly IC 1S also useruI TOF restoring calis tnat were previously displaced due to high error rate or congestion if either of the Error Monitoring or Congestion Monitoring features is in use on the primary back onto their primary route Note that any calls that fail to re establish down a better trunk will simply be remade on the trunk from which they were cleared 4 3 1 5 Congestion Monitoring Parameters The Congestion Monitoring feature allows some control over the sharing of bandwidth between calls using a trunk and is configured by means of these parameters It can be arranged that during periods of congestion lower priority calls will be successively re routed via alternative trunk ports allowing greater bandwidth for the higher priority calls If it is believed that congestion may be occurring within the network then the task of locating and diagnosing the problem can be helped by looking at the Utilisation levels measured at the trunk ports The most recent utilisation measurements can be readily inspected on the Configuration Node Configuration Detailed Link Status Display Screen A full description and configuration guidance for Congestion Monitoring and Control can be found in Appendix M 4 3
122. ed while you are logged onto an application then the Node Manager will ensure that your connection to the application is also cleared X890 304751 Issue 1 2 9 Rev 0 X890 304751 Issue 1 2 10 Rev 0 3 Configuring a Single PSE The PSE can operate on its own or as part of a network of PSEs This chapter concentrates on the concepts needed to configure a single node but also briefly refers to inter node trunks where applicable For full details on the latter refer to Chapter 4 3 1 Physical and Logical Links and Ports 3 1 1 Physical Links and Ports When operating as a single node the PSE simply routes connections between the various devices connected to it These can be e External packet mode devices such as PADs X 25 card equipped PCs X 25 capable hosts gateways to other X 25 networks etc e Xpress Applications i e internal packet mode devices In both cases the connection between the PSE and the attached packet mode device is called a link This link carries the traffic between the PSE and the device according to the X 25 protocols see Appendix A For the purposes of this section it will be assumed that for external devices this link is provided by a direct physical connection such as a simple piece of cable or digital leased line etc Other connection possibilities will be explored later In the case of an Xpress Application the link is provided by the internal software equivalent of a piece of cable and application lin
123. el software The user accesses the channels by making an X 25 call from a Triple X PAD to one special network address for standard input output and another network address for standard error Signals A signal is an interrupt to an application process Signals may be raised for different reasons such as arrival of X 25 out of band data or at the request of another process Similarly the effect of signals varies from termination of the interrupted process to the invoking of a handler routing defined by the application Environment Variables Applications may read the values of system variables using the getenv system call e g Nodenumber of the PSE Number of the slot on which the application resides Logical Port Number assigned to the application link By using other system calls the application can read the time and date as maintained by the Node Manager J 5 3 Management Services Node Management Services The PSE Node Manager allows the operator to install and select applications It also re loads applications when necessary and stores core dumps to assist with debugging An application can raise Event messages which the PSE Node Manager will handle invoke file operations on the PSE Node Manager s floppy hard disk system The PSE Node Manage can interrogate the Xpress Kernel software to provide information about the use which an application is making of the Xpress X 25 service The Xpress Kernel softwar
124. em will not boot up 2 Switch on the power The system will now undergo a series of hardware diagnostics If these are successful the manager will then proceed to boot up itself and the rest of the system This process takes approximately 5 10 minutes depending on the number and type of boards 3 Press the RETURN key on the manager terminal and it will display the Mini PAD prompt The printer if you have one attached to the system will output events indicating the status of the boot up The PSE is now operational Having initially booted up the system users are strongly advised to make working copies of all floppy diskettes supplied with the equipment using the Copy Disk manager facility described in Section 5 3 of the User Guide and to keep the master diskettes in a secure place It is good practice to make regular back up copies of all operational diskettes in this Way 2 1 2 What the PSE Does to Boot Itself Up The indicator on each board provides a running commentary on what is taking place during the boot up sequence Stage 1 The hardware is tested When the system is first powered on all the boards indicate t This means that hardware diagnostics are being performed If the diagnostics fail an error code is flashed up on the card s X890 304751 Issue 1 2 1 Rev 0 aispray panel 1 NE coae comes In two parts tne test number ana the failure code The characters delimit the code The exact syntax is lt test
125. ement Centre Network Management System An Xpress PSE Outgoing Called calling Address Translation Open Systems Interconnection Packet Assembler Disassembler Public Data Network Packet Switch Exchange Packet Switched Private Data Network Packet Switched Public Data Network BritishT elecom Packet SwitchStream Power Supply Unit Permanent Virtual Circuit X 25 Level 3 Qualified data bit Random Access Memory Read Only Memory Recognised Private Operating Agency 8325 equivalent of SP XIM also called XSAC Six Port Access Module a type of ACM X890 304751 Issue 1 H 3 Rev 0 SDLC SNA SP XIM SVC TNIC Trunk Trunk Group UM UPM VDU V 11 V 24 V 35 V 36 V 54 VC X 2 X 3 X890 304751 Issue 1 H 4 Synchronous Data Link Control IBM Systems Network Architecture Six Port X 25 Interface Module aUPM SAM card combination Switched Virtual Circuit Transit Network Identification Code An inter node link between two Xpress PSEs A Hunt Group comprising Xpress trunks Calls are distributed across the members of a Trunk Group as for a Hunt Group Utility Module a type of ACM Universal Processor Module Visual Display Unit CCITT Recommendation which concerns electrical characteristics for balanced double current interchange circuits for general use with integrated circuit equipment CCITT Recommendation which concerns definitions for interchange circuits between DTE and DCE i e modem CCITT Recommenda
126. en configuring a node first allocate logical ports to all the required physical ports using the Configuration Logical Port Allocation menu 1 Select the create new logical port screen and enter the logical port number to be allocated 2 Enter the bay slot and link numbers of the physical port together optionally with a short description of the port 3 Press PF1 to submit the form Repeat this process for each port Logical port allocations for unused physical ports can be carried out later when needed Options also exist to edit and delete ports but before you can do this you must take the port s Out of Service by using the Configuration Port Configuration X 25 or Trunk port Configuration change state of port Screen A list of ports can be obtained in numerical order by using the Logical physical port display screen and in physical order by using the Physical logical port display Screen Logical port numbers are assigned specific meanings Range Usage 0000 to 6999 Logical Port see Section 3 4 7000 to 7999 Logical Port This range is reserved for Application Links 8000 to 8999 Hunt Groups see Section 3 8 9000 to 9999 Virtual DTEs automatically created see Chapter 6 T0000 to T6999 Trunk Port See Section 4 3 T7000 to T7999 Trunk Port In a future version this range may be used for Application Trunks T8000 to T8999 Trunk Hunt Groups see Section 3 8 T9000 to T9999 Trunk Port see Section 4 3 Befor
127. en the Internal Address is used for routing A called address that does not match any entry is passed on for routing as an Internal Address e Match Address Wildcard characters supported n matches any digit 012 matches any one digit in a set A range of digits may be specified e g 0 7 matches any digit from 0 to 7 matches zero or more occurrences of the preceding digit allows a portion of the match address to be tagged and referred toin the internal address Space allowed for clarity Match Addresses may be up to 15 digits long e Internal Address Special characters allowed 1 2 9 tag parameters 1 matches the first tagged portion of the match address 2 matches the second tagged portion of the match address and so on X890 304751 Issue 1 4 18 Rev 0 Impress the current node space allowed for clarity Internal address may be up to 14 digits in length e Example Address Analysis Table Match Address 1100 nnn 9 nnn n 1234 112 000 1 4 1234 112 002 4 9 1234 56n 1200 23n 9876 nnn nnnn 7639 n 1100 Inn nnnn n 1100 nnn nnnn n n or Se a er O opm Mapping will occur as follows Internal Address 1100 1 9 2 3 1100 001 1 1100 002 1 1100 013 1100 8001 1100 1 2 2342 1100 100 1100 1 NULL a The Address Analysis table has great scope for locking out access to the Xpress Manager so this entry allows access to the Node Manager and
128. erefore 32 times the number of X 25 cards in the node i e 128 clients per 8325 256 clients per 8425 and 512 clients per 8525 G 5 2 Buffering Each server will buffer a fixed number of data packets of any length before it starts to discard any data received over its host call For servers running in slots containing 1 MByte UPM1 processor cards this limit is 64 data packets for all other servers the limit is 160 data packets For every client there is a queue on which data is buffered if necessary This means that a client who is not acknowledging data fast enough will lose data if the particular buffer queue fills up This will not however cause any other client tolose data The available buffer memory is shared equally between all the active clients buffer queues It should be noted that if large packets are being used on the host call 512 4096 bytes or there is a lot of non ABS traffic on a card then it is possible that the card will go into slowdown mode and flow control the host call automatically If the affected server is being fed by a higher order server X890 304751 Issue 1 G 16 Rev 0 IN ahnrerarcny then tne nigner oraer server will Oulrer Gata aS GeScr oea above treating this server the same as any other client If data is discarded by any server then events will be generated as described in Section G 6 G 5 3 Throughput The ABS is capable of broadcasting host data being received at the rate of 1 128
129. es a call to 1100 140 9079 005 If the PAD is directly connected to an X 25 port on the card in node 140 slot 7 the client could call 1100 140 9999 005 using the slot 99 convention meaning this slot Once the client is connected to the server any host data received by that server will be duplicated and forwarded to the client Host data includes data with the Q bit set for example X 29 messages These messages are duplicated and forwarded as if they were normal user data It should be noted that any responses to these messages sent by the clients will be discarded by the server Any data sent by the client will be acknowledged and discarded Any interrupt packets sent by the client will also be acknowledged and discarded If the client sends a reset packet this will be acknowledged and will cause any locally buffered data to be discarded The host call and any other client calls will be unaffected in all cases If the client s terminal flow controls the server or is running too slowly to accept the data stream from the server for example the host sends data at the rate of 100 characters per second then data will be buffered as described in Section G 5 and will then be discarded once the buffer space is full There is no mechanism whereby any client can flow control the host This is to prevent a situation whereby all clients are limited tothe reception rate of the slowest client which can be zero in the case of a flow controlled te
130. ess Analysis Table enables non Xpress addressing schemes to be supported A network of Xpress PSEs can be inserted transparently into an existing non Xpress network without any change to any existing addressing scheme e Address Translation The Xpress Address Translation tables operate on the fringes of the network to provide policing and conversion of User Addresses e Gateways An Xpress network can provide many gateways to PDNs or private networks These gateways can be distributed across different nodes in an Xpress network X890 304751 Issue 1 4 1 Rev 0 e Closed User Groups CUGs An Xpress network can support up to 65535 CUGs The PSEs map between the different indices so that a CUG appears to have different CUG numbers on different nodes e PVCs PVCs provide a permanent channel across the network between two X 25 ports This allows the ports to communicate via the PVC with no need for a call setup procedure e Automatic Rerouting On Trunk Failure If a trunk or intermediate node fails within an Xpress network or a trunk gets automatically closed due to high error rates then the PSEs attempt to reroute affected SVCs or PVCs by using any alternative trunks within that network This procedure is carried out automatically by the PSEs By default the PSE discards any data in transit within the network when a call is rerouted However if Network Data Integrity is enabled at the originating X 25 port or if a call uses D
131. evel Note that X 25 X 75 Application port configuration Could be used in place of Trunk port configuration at this point as the two are identical at the physical and data link levels Type t70 RETURN to select logical port T0070 Type c RETURN 10 RETURN to select 256000 bps from the clocking speed options box Type p RETURN x21 RETURN to select the X 21 physical interface Submit the screen with PF1 Type PF3 f RETURN t70 RETURN to go to the frame relay core level configuration screen for port T0070 Type rb RETURN y RETURN to enable R bit support Submit the screen with PF1 20 21 xw o 22 23 24 xw 25 26 27 Type PF3 to return to the Trunk port configuration menu Step 4 Configure the two Xpress Trunk logical ports and map them onto the frame relay physical port 28 Typed RETURN t70 RETURN to configure the data link level of port T0070 29 Typed RETURN 16 RETURN to assign this logical port and hence virtual physical port 0 7 0 to DLCI 16 on the frame relay physical port X890 304751 Issue 1 4 12 Rev 0 Submit the screen with PF 1 31 Typer RETURN t71 RETURN to configure logical port T0071 Repeat steps 29 and 30 but use a DLCI of 17 Type PF3 n RETURN t70 RETURN to configure the network level of port T0070 34 Type L RETURN 1 RETURN u RETURN 128 RETURN to set the lower and upper logical channel range boundar
132. eway 20 0020 Secondary gateway 20 0021 Tertiary gateway 31 0024 Calls specifying that DNIC are routed over the network to the first available gateway port Apart from the DNIC the address is not examined further and is transported transparently In order for such a call to traverse a network of PSEs successfully the gateway routing must be configured at every node en route to the gateway port The call is attempted to each of the gateway ports in order of priority until it is successfully connected Gateway ports are identified by their node number and logical port number which may be that of a hunt group This allows a number of gateway ports on a single node to be collected together and calls balanced over all the ports 3 9 3 Calls From a PDN Calls coming into an Xpress node from a foreign network will contain addresses specific to that network These addresses may be mapped to an Xpress PSE network address by using the PSE s address translation tables or by using the Address Analysis Table X890 304751 Issue 1 3 45 Rev 0 3 9 4 Reserved DNICs The following DNICs are reserved DNIC Usage 1100 Xpress Internal Addressing Scheme for Normal User Calls Call Re establishment Centralised Printing PVCs 9990 to 9999 DTE Clear is not fatal at call setup time 9990 to 9997 Reserved for future use 9998 Multiple ACS see Appendix E 9999 Used by the PSE to make management calls to Cray Network Management Centres e g to report event
133. f UNIX library calls e g printf scanf The UNIX socket protocol is used to implement the communications services provided by Xpress J 55 2 Applications Environment Xpress emulates the following aspects of the UNIX environment for applications Processes and Process IDs Each applications task which uses the Xpress system library calls must have a unique process ID Applications achieve this by invoking the newthread system call whenever they create a new task which uses the Kernel functions Files Sockets and File Descriptors File descriptors are numbers which are used by processes to manipulate system resources such as disk files and communications sockets Applications use sockets to access the communications services provided by Xpress An application process must create a new socket using the socket system call each time it wishes to use a communications service File descriptors are simply integers which are passed as parameters to system calls File descriptors 0 1 and 2 are reserved for Standard Input Standard Output and Standard Error respectively Standard Input Output and Error Xpress provides a separate set of standard input output error channels for each instance of an application The applications software accesses these channels by means of the library calls such as printf X890 304751 Issue 1 J 6 Rev 0 The standard input output and error channels are maintained by the Xpress Kern
134. f another Xpress PSE This is similar to the above method In order to connect to the Node Manager of node nnn type REMOTE nnn RETURN or R nnn RETURN after the Mini PAD prompt has been displayed e From a Cray Network Management Centre The Cray NMC Operator Guide provides details of transparent management of the PSE e From a VT100 terminal through a PAD The connection address for transparent logon is 1100nnn9000 where nnn is the number of the node being logged onto Once the call is connected the Triple X PAD will have its parameters changed by the Node Manager After the call is cleared the default PAD profile will be restored The affected parameters are described below X 3 Parameter Value Purpose 2 0 Local echo 3 2 Data forwarding character 4 1 Data forwarding timeout X890 304751 Issue 1 2 3 Rev 0 U LIne TOLAING 13 0 LF insertion 15 0 Editing In all these access methods while the node is managed the time display is updated every minute Logging out may be performed at the main menu typing L RETURN or by clearing the X 25 call The VT 100 should be configured for 8 bits no parity 2 2 2 Using the Mini PAD Once the PSE has successfully powered up see Section 2 1 you can activate the Mini PAD by pressing RETURN on the terminal The Mini PAD uses RETURN to work out the speed and parity settings of the terminal and it may be necessary to press the key two or three times for it to dothis If pres
135. fines the interface between two Signalling Terminal Equipments STEs STE X is considered to be the STE of this network and STE Y is considered to be the STE of the other network An X 75 interface i e an STE X STE Y interface is sometimes abbreviated to the X Y interface Depending on the configuration one STE operates as an X 25 DTE and the other STE as an X 25 DCE In fact Xpress does not use the term STE for an X 75 interface but instead uses the X 25 terms DTE and DCE On a per call basis Xpress appears as different types of X 75 networks Xpress operates as an originating network for a call made from an Xpress X 25 port to an X 75 port Xpress operates as a transit network for a call made from one Xpress X 75 port to another X 75 port Xpress operates as a destination network for a call made from an Xpress X 75 port to an X 25 port Xpress assumes that X 75 calls use full X 121 addressing in that the first four digits of the address area DNIC An X 75 call can be routed over a mixed Xpress network provided that all nodes run Version 3 software or later and the X 75 call set up packets contain at most one group of X 25 facilities Xpress supports both SVCs and PVCs at X 75 interfaces however no special X 75 support is provided for PVCs Level 1 the Physical Layer Xpress supports Level 1 of X 75 interfaces in exactly the same way as for X 25 interfaces see A 4 1 X890 304751 Issue
136. g 3 4 8 Configuration Procedure 3 5 Virtual Circuits 3 5 1 Xpress Internal Addressing 3 5 2 SVC Configuration 3 5 3 PVC Configuration 3 5 4 PVC Call Establishment 3 6 Examples of Port Configuration 3 6 1 Example 1 X 25 Port Configuration 3 6 2 Example 2 Application Port Configuration 3 7 Curing Problems 3 7 1 Node Does Not Power Up 3 7 2 XIMs Not Loaded 3 7 3 Errors During Configuration 3 7 4 X 25 Data Link Down 3 7 5 X 25 Call Failed 3 7 6 Failed Installation of an Application 3 8 Hunt Groups 3 8 1 Hunt Group Addressing 3 8 2 Call Distribution within a Hunt Group 3 8 3 Trunk Groups 3 9 X 25 X 75 Gateways 3 9 1 Internetworking DNIC IDNIC 3 9 2 Calls Toa PDN 3 9 3 Calls From a PDN 3 9 4 Reserved DNICs 3 9 5 DNIC Barring Table DBT 3 9 6 X 25 Gateway 3 9 7 X 75 Gateway 3 10 Frame Relay Gateways 4 Configuring a Network 4 1 ntroduction 4 2 Node Numbering 4 3 Trunks X890 304751 Issue 1 0 6 0 10 3 12 3 14 3 16 3 18 3 22 3 22 3 24 3 27 3 27 3 28 3 28 3 29 3 30 3 30 3 35 3 38 3 38 3 38 3 39 3 39 3 39 3 40 3 42 3 42 3 42 3 43 3 44 3 44 3 45 3 45 3 46 3 46 3 47 3 48 3 49 4 1 4 1 4 3 4 4 Rev 0 4 5 L trunk Fort Contguration 4 3 2 Trunks over Frame Relay 4 4 Routing 4 4 1 The Routing Algorithm 4 4 2 Routing Procedure 4 4 3 The Routing Table 4 4 4 Routing the Call 4 4 5 Using Trunk Groups in the Routing Table 4 4 6 Hop Counts 4 5 Addressing 4 5 1 Address Analysis 4 5 2 Address Tra
137. g is required to operate e Priority Class There are 4 priority classes numbered 1 to 4 Enter the appropriate number to select a row in thetable Then select one of the three threshold parameters to modify for the selected priority class These are described below e Refuse at This parameter sets the call refusal threshold for a given priority class Whilst the measured utilisation at the port equals or exceeds this value the port will disallow further calls of this priority to be established through the port Existing calls are unaffected the refuse threshold only affects calls in the process of being established The default value of 100 utilisation allows all calls and so effectively disables the refusal feature e Clear at This parameter sets the call clearing bumping threshold for the selected priority class The port software starts clearing other calls to free up bandwidth for this class of call whenever thereis at least one established call having this priority the measured utilisation exceeds this threshold value and X890 304751 Issue 1 M 4 Rev 0 there are some suitable calls that can be cleared The choice of which class es of calls are candidates for being cleared is made by setting the Priority to Clear parameter below Call clearing continues until the utilisation level subsides below the threshold or the number of calls present at this priority goes to zero The default value of 100 d
138. gits of the client whose data is being discarded NNNN is the number of packets which have been discarded since the call was made or was last reset by the client and SS represents the slot number of the server doing the discarding X890 304751 Issue 1 G 18 Rev 0 Appendix H AAT ACM ACS ANSI Application ASCII Async BCD BECN bps BSC Cause Code CCITT CNIC Co resident Application CRC CUG D bit Glossary Address Analysis Table Application Connector Module e g UM XIM1 Access Control Server American National Standards Institute Application software which provides functionality and physical interfaces that are additional to those provided by Xpress An application is loaded with the Xpress Kernel onto a card combination American Standard Code for Information Interchange Asynchronous Binary Coded Decimal Backward Explicit Congestion Notification Bits per second IBM Bisynchronous protocol This is carried by X 25 clear reset and restart packets to indicate the reason why a call has been cleared reset or a link restarted Consultative Committee for International Telegraphy and Telephony Clearing Network Identification Code An application which is co resident on a UPM with the Xpress Kernel Cyclic Redundancy Check Closed User Group X 25 Level 3 Delivery confirmation indication bit X890 304751 Issue 1 H 1 Rev 0 DCE DCX DBT DE bit Distribution disk DLCI DNIC DTE FCS F
139. gure and modify and less likely to introduce delays caused by data passing through multiple servers Both the examples in Figures G 2 and G 3 assume that all the servers are co located on node 1 In reality it is possible to configure multiple servers on a number of nodes all that is required are the requisite AAT entries and standard inter node trunks and routing See Figure G 4 X890 304751 Issue 1 G 8 Rev 0 Host port 40 Server 1 Server 2 Server 3 Node1 slot 1 Node1 slot 2 Node1 slot 3 up to 32 clients up to 31 clients up to 31 clients Note connecting lines denote X 25 calls not physical connections AAT Entry Match Address Internal Address 1100 001 9019 995 1100 001 9029 005 1100 001 9029 995 1100 001 9039 005 1100 001 9039 995 1100 001 0040 000 Figure G 2 Example of a Linear Multiple Server X890 304751 Issue 1 G 9 Rev 0 Host port 40 Server 3 Node1 slot 3 Server 2 Node1 slot 2 Server 1 Node1 slot 1 up to 32 clients up to 32 clients up to 32 clients Note connecting lines denote X 25 calls not physical connections AAT Entry Match Address Internal Address 1100 001 9019 995 1100 001 9029 005 1100 001 9029 995 1100 001 0040 000 1100 001 9039 995 1100 001 9029 005 Figure G 3 Example of a Hierarchical Multiple Server Note that in both cases the fact that the host call is shown to be connected toa particular server does not mean that the host has to be physica
140. h allows the called device to clear a call in direct response to a call request specifying an alternative network address to which the call should be deflected The default is Disabled Note Call Deflection only works when the port is On Line e Call Deflection Data Transfer Allowed X 25 only This facility is a non standard extension to the Call Deflection facility which allows the connected device to deflect an established call i e a call which is in data transfer state It is provided to support Cray Access Control Server ACS See Appendix E for details The default is Disabled e Call Deflection Referral Enabled X 25 only This non standard facility allows an unsuccessfully deflected call to be referred back to the original called port See Appendix E for details The default is Disabled e Call Redirection Enabled Alternate Network Address This is the X 25 1984 facility There are two types of Call Redirection Systematic and Incidental If Systematic Call Redirection is enabled on a port the PSE will redirect all calls destined for that port to an alternative port as specified by the Alternate Network Address X890 304751 Issue 1 3 20 Rev 0 parameter Ne originally aaaressed port will not receive any Calls Incidental Call Redirection works in the same way except that the call is sent first to the original destination and only redirected if it fails for some reason The Default is Disabled The Alternate
141. he Expand command allows you to display details of an application which is either already installed onto the PSE or present on a distribution disk When you select the Expand command you will be prompted to specify the disk drive If you specify drive A then you will be prompted for the entry number of the application If you specify drive B then the Node Manager will select the single application on the distribution disk The details of the application are displayed on the screen If necessary the list of card types and load files will continue onto a second line You can exit from the Expand command by pressing RETURN Note that the utility shows all the required load files even the Xpress Kernel load file which is distributed on the Xpress disk set and not on application distribution disks 5 5 5 Background Information This section contains background information about how the above commands operate The Install command supports only one application per distribution disk X890 304751 Issue 1 5 12 Rev 0 mowever the aistripution QiSkK May Noid Several DUIIGS OF tNe application i e load files for different types of cards In such a case a single invocation of an Install Delete or Expand command operates on all the different builds of the application The Xpress Kernel software is issued on the Xpress disk set It is not present on application distribution disks The database file on the application distribution disk must
142. he new boot filename Hence when you copy the new boot file onto the disk it takes up its correct place on the disk since it is overwriting the area reserved for the old file You should also check that the file sizes of the new and old boot files are the same If they are not identical stop immediate y and contact Cray support X890 304751 Issue 1 L 8 Rev 0 L 4 Software Licensing The number of nodes on which you are authorised to load a specific version of the Xpress software is governed by the terms of your licence agreement Be aware of this when copying software from node to node X890 304751 Issue 1 L 9 Rev 0 X890 304751 Issue 1 L 10 Rev 0 Appendix M Congestion Monitoring and Control M 1 Introduction The Congestion Monitoring and Control feature can be used to improve the performance of an Xpress network as it allows each node to take into account the degree of congestion detected at its ports when making its routing decisions The mechanism used embodies the concept of call prioritisation for differentiating between different traffic types The feature operates on congested trunk or X 25 X 75 ports and works by diverting low priority calls on to alternative routes to increase the bandwidth available for higher priority calls The Packet Switch uses its re routing and transparent call re establishment capabilities to find alternative paths for the affected calls Congestion Monitoring is typically used on a trun
143. heir slot number The frequency at which this report is printed is governed by the statistics reporting interval INCS statistics cannot be viewed dynamically For each module configured into the report the following details are produced module type e g XIM1 module state e g Operational connection level statistics task level statisticsupm level statistics bus level statistics These statistics are for Cray use only 5 10 11 Set up Reporting Interval The statistics control report screen lets you change the frequency at which the reports are produced All three reports are produced at the same frequency The interval can be set for intervals from 15 minutes up to 28 days If long reporting intervals are used some statistics counters may wrap around X890 304751 Issue 1 5 23 Rev 0 wnicn may give misieading results OdlSabdle Statistics colection on the node set the reporting interval to 0 To change the content of a report use the Statistics Modify Report option X890 304751 Issue 1 5 24 Rev 0 5 11 Status Displays Four options are available for viewing the status of the node These are e Display node status e Detailed Link display e Display link circuits e Summary link display It is possible for things to change for example calls may clear even as the data for this screen is being collected Therefore the picture presented may not be completely accurate Note that no explicit reference to frame relay l
144. hen rename it with the move command after logging into the remote node to perform the move Transfer the new file s to remote drive b and then do a local file copy from drive b to drivea of the appropriate files after logging into the remote node L 2 2 Security Violations It is possible for a System M anager to copy the access rights configuration file from one node to another The access rights on the two nodes might differ considerably for various reasons It is vitally important to ensure that the passwords and access rights of every node in the network are correctly set up A single node with weak passwords or access rights would allow a determined rogue System Manager to circumvent the access rights and passwords of the other nodes in the network by simply overwriting the appropriate file on the remote node X890 304751 Issue 1 L 3 Rev 0 L 3 Example Operations L 3 1 Configuration File Backups Assume the following Node number 1 is a central node A remote node number 30 contains a series of configuration files you would like to back up You arerunning Version 8 2 on all nodes The operator is working from the Mini Pad at node 1 1 Put a blank disk in drive b of node 1 Login to node 1 and format the disk Logout of node 1 Login to node 30 UO BF W N From the Utilities Disk Utilities File Copy screen copy all config files from the local drive a to the remote drive b of node 1 The from file will be
145. hich allows the exchange of 128 bytes of user data between two DTEs without the need to establish an SVC as explained below 1 The DTE sends to its DCE a Call Request packet requesting the Fast Select facility and holding up to 128 bytes of user data 2 The packet is forwarded to the called DTE which depending on the circumstances may either respond with a Call Accepted packet to establish an SVC or respond with a Clear Request packet holding 128 bytes of user data 3 In the latter case the user data in the Clear Request packet is passed to the calling DTE which responds with a Clear Confirmation packet to complete the removal of the SVC X890 304751 Issue 1 A 8 Rev 0 A 2 4 X 25 User Facilities Supported by Xpress PSEs Tables A 1 and A 2 list the User Facilities indexed by X 2 1988 numbers Xpress supports all the User Facilities which X 2 specifies as being essential as well as most of the optional facilities The supported by Xpress column refers to whether the relevant port is set to 1980 CCITT or 1984 88 CCITT at the network level Supported by Xpress X 2 1988 User Facility 1980 1984 88 Extended frame sequence numbering Ne O Multilink procedure Not supported Not supported On line facility registration Not supported Not supported Extended packet sequence numbering modulo 128 D bit modification Packet retransmission Incoming calls barred No Yes Yes Yes Yes Outgoing calls barred Yes
146. host data to be on subaddress 000 It does not take any notice of the remainder of the called address Host slelb ele Ei ES lalla a Physical Connection Broadcast Data Figure G 1 Example of Single Node Single Server ABS Configuration There are six clients connected via three PADs which arein turn connected to ports 25 50 and 54 It has been decided to use the ABS server in slot 4 this is an arbitrary choice any slot could be used All the terminals make calls to address 1100 001 9049 005 to pick up the broadcast data X890 304751 Issue 1 G 6 Rev 0 The ABS server makes its host call on 1100 001 9049 995 which is mapped by an AAT entry to 1100 001 0030 000 to route the host call to port 30 As the host address actually received by the host computer would be the original server called address 1100 001 9049 995 an outgoing called address translation on port 30 is needed to replace the 995 subaddress with 000 as required by the host It should be noted that if a host call is destined for a port on a node other than the one on which the server making the call resides then it will be necessary to configure an AAT entry on each transit node to map the fixed host address 1100 nnn 9ss9 995 to the required host network address This is because the AAT mechanism leaves the original called address in the call request packet and it is not possible to carry out address translations at trunk ports or vir
147. how to install an application and then make a call to it from one of the PADs This example assumes that slot 3 of the node holds a card combination on which an application can run It assumes that the application is co resident i e it resides on the UPM3 together with the Xpress Kernel software It is also assumed that the application has an MMI which can be accessed at sub address 01 1 2 Log on to the Node Manager by entering L RETURN your user name and password Note that the password will not be echoed Select the Configuration Node Configuration node status display screen This should show the application card which may bein Unknown state unless it happens to be suitable for running the Xpress X 25 software in which case it should be in Operational state Press PF4 to return to the main menu Now install the application onto the PSE Select the Utilities Install Delete Expand Application screen The screen will list all the applications which have already been installed onto the PSE Note that the list includes native applications which are the Xpress Node Manager X 25 and dumper software which are always present Insert the distribution disk which holds the application software into drive B Type i RETURN RETURN PF 1 to install the application The screen should display the message Application installed successfully If it isn t refer to Section 3 7 6 The screen should now list th
148. ial configuration X890 304751 Issue 1 E 9 Rev 0 X890 304751 Issue 1 E 10 Rev 0 Appendix F V 54 Modem Test Facilities F 1 Introduction This appendix provides a brief description of the modem test facilities provided by the PSE The PSE can set up both local and remote modem loops It can also generate test messages to the modems and check the returned data for errors Use of these test facilities disrupts the normal operation of the port concerned Therefore the port must first be put out of service The loopbacks only operate at a port with a V 24 physical interface and connected to a modem that supports V 54 loopback procedures A special cable is needed to connect the PSE port to the modem X890 304751 Issue 1 F 1 Rev 0 F 2 Modem Test Loops The PSE supports V 54 Test Loop 2 remote loopback and Test Loop 3 local loopback See Figure F 1 below Loop 3 Loop 2 Figure F 1 Modem Test Loops Local Loopback A local loopback allows the data path between the PSE and Modem A to be checked Remote Loopback A remote loopback allows the data path between the PSE and Modem B to be checked Test Loops may be enabled and disabled via the node manager X 25 Port Trunk Physical Level Configuration screen A port s loopback state is not saved on disk so if a PSE is re started the test loops on all of its ports are initially disabled The status of the modem test loops can be examined via the node manager X 25 Port Trunk
149. ialling This is not mandatory as it is possible to arrange one way operation by setting only one end of a trunk to be dial up in which case the other end will happily accept incoming calls but will not be able to initiate calls When making a dial up call the software will hold onto the X 25 call that initiated the dialling sequence plus any others which follow for a user definable length of time the dial up timeout after which the call attempt will be abandoned and the X 25 call s will be cleared back for possible rerouting if the link has failed to come up In the case of a symmetrically configured trunk carrying X 25 calls in both directions backed up by a dial up link the software automatically copes with the possibility of a dial collision This can occur when the two ends of the trunk realise it has gone down at the same time and simultaneously clear back and reroute the X 25 calls across the dial up link If this occurs the dial up software will delay the calls at one end of the dial up link to give the other end the chance to make the call and have it answered K 2 2 V 24 Interface Circuits The signalling used to drive a V 24 dial up interface is based on V 25bis control signals as detailed below DTR Pin 20 circuit 108 1 Connect Data Set to Line Raised by the node to signal that the attached DCE should dial a pre defined number to establish a connection Dropped by the node to signal that the connection should be dr
150. ics 5 10 7 Modify Report 5 10 8 Link Statistics Report 5 10 9 Module Statistics 5 10 10 Intra node Communications Subsystem I NCS Statistics 5 10 11 Set up Reporting Interval Status Displays 5 11 1 Display Node Status 5 11 2 Detailed Link Display 5 11 3 Display Link Circuits 5 11 4 Summary Link Display Diagnostics and Error Handling The Virtual DTE Facility 6 1 1 What It Is Used for X890 304751 Issue 1 0 8 J 5 9 5 10 5 10 5 10 5 11 5 11 5 11 5 12 5 12 5 12 5 13 5 14 5 15 5 16 5 17 5 18 5 18 5 19 5 20 5 20 5 20 5 21 5 21 5 21 5 22 5 22 5 22 5 22 5 23 5 23 5 25 5 25 5 26 5 26 5 27 6 1 6 1 6 1 Rev 0 0 1 2 MOW ItIS ACCessed 0 1 6 1 3 When It Should be Used 6 2 6 1 4 Node Manager Virtual DTEs 6 2 6 2 Module Crashes 6 4 6 3 Dump Files 6 5 Appendices A X 25 Frame Relay and Packet Switching A 1 A 1 Introduction A 1 A 2 The X 25 Recommendation A 2 A 2 1 Other Standards Relevant to X 25 A 2 A 2 2 How the X 25 Protocol Works A 3 A 2 3 Procedure for a Switched Virtual Circuit A 7 A 2 4 X 25 User Facilities Supported by Xpress PSES A 9 A 2 5 Additional Notes about Xpress Support of Some A 12 X 25 Facilities A 2 6 Calls Between X 25 1980 and X 25 1984 1988 A 15 Ports A 2 7 Xpress and the X 75 Recommendation A 15 A 3 FrameRelay A 22 A 3 1 Introduction A 22 A 3 2 How the Frame Relay Protocol Works A 22 A 3 3 Series 8000 PSEs and Frame Relay A 25 B Error Causes and Diagnos
151. ies 35 Submit the form with PF 1 36 The configuration of the network level of logical port T0071 is identical so type k RETURN t71 RETURN PF 1 to configure it Step 5 Place the ports on line 37 Type PF4 and select Configuration Module Configuration Change module link states 38 Type 7 RETURN on RETURN PF1 to put both ports on line That completes the configuration of the twotrunks Tothe rest of the software these trunks are now indistinguishable from normal non frame relay trunks For example they can be used in the routing and PDN gateway tables in exactly the same way as any other trunk X890 304751 Issue 1 4 13 Rev 0 4 4 Routing 4 4 1 The Routing Algorithm The Xpress PSEs use a static routing algorithm to guide calls through an Xpress network Each PSE contains a routing table that specifies the route that must be taken in order to forward a call to any other node in the same network When a PSE has to forward a call to a remote node it looks up that node s entry in the routing table which indicates the trunk over which it should forward the call Normally up to three routes are configured to a given node the primary secondary and tertiary routes The call is attempted along each route in turn until either the call is successfully connected or no more routes remain It is the user s responsibility to configure and maintain the routing table 4 4 2 Routing Procedure The addresses in a
152. if the terminal is powered off 2 2 8 Special Characters Backspace or Delete CTRL U RETURN ESC X890 304751 Issue 1 Deletes a character Deletes a line Causes the previous characters to be input Deletes line and starts the function key sequence i e PF1 PF2 PF3 and PF4 2 7 Rev 0 2 3 Management of Applications Once you are logged onto the Node Manager you can select the Manage applications menu option to access the management screens of applications which have been installed onto the PSE The Application Management screen lists the application managers which have been configured When first entered the screen displays the first page of the list Use the Next page command to view the next page of the list the Prev page command to view the previous page and the First page command to view the first page The Application Management screen operates in the same way irrespective of whether you are locally or remotely logged into the PSE Node Manager 2 3 1 Configuration of Applications Management The Application Management screen provides the Add Edit and Delete commands with which you can add new applications to the list edit existing entries and delete unwanted entries The Node Manager stores the information on disk To add an entry tothe list you must supply the following information Description free format text to provide a meaningful description of the application Address The netwo
153. ile 5 3 5 M ove file 5 3 6 Du mp utilities 5 4 amp 6 3 D elete dump file 5 4 1 P rint dump file 5 4 2 I nstall Delete Expand applications 5 5 P rint utilities 5 6 R outing specification Po rt configuration Pr inter configuration M odule configuration L ogical port allocation MJanage applications 2 3 L ogout X890 304751 Issue 1 l 4 Rev 0 Appendixj Xpress PSE Applications This appendix gives an overview of the support which Xpress software provides for applications software J 1 Overview J 1 1 Native Applications Native applications are distributed as an intrinsic part of the Xpress communications and management software The Xpress native applications are the Node Manager Dumper and X 25 X 75 communications software J 1 2 Lodger Cards Lodger cards are plugged into and draw power from an Xpress PSE The applications which run on lodger cards do not interface to the Xpress Software and must provide all their own management and communications services J 1 3 Imported Applications mported applications software resides on an intelligent ACM attached toa UPM or co resides with the Xpress software on a UPM The latter type of application is called a co resident application Imported applications are not distributed as part of the Xpress software but they do interface to the Xpress Kernel UPM software The Xpress Kernel software provides access to the Xpress communications and management services
154. ill determine firstly if there are any calls which are candidates for X890 304751 Issue 1 M 5 Rev 0 cearing anda will then crear the number of calis specitied Dy Calls Cleared in one CCI This progressive clearing means that the system is able to reassess the utilisation level as the number of calls decreases and prevents an overreaction in which toomany might be cleared at once As soon as the utilisation subsides below all the applicable clear thresholds the clearing process stops It is advisable to set the Call Clearing Interval to at least 30 seconds i e twice as long as the congestion monitoring period which is fixed at 15 seconds This gives the system time to assess the effect of the clearing activity it has just completed before commencing to do any more The default value of 1 minute is appropriate in most cases Valid settings for Call Clearing Interval are in the range 15 seconds congestion monitoring period to 10 minutes Valid range for Calls Cleared in one CCI is 1 100 e Priority Class Profile This parameter does not normally apply on the congested trunk itself and can be left at its default value It should be configured on the X 25 X 75 link ports at which calls enter the Xpress network and is discussed in the next section M 2 2 Configuring X 25 X 75 Link Ports This section covers configuration of ports that connect the Xpress network to user equipment and gateway ports to other networks when Congestion Mon
155. in the same CUG A does not have Outgoing Access D BandD arenot in the same CUG B does not have Outgoing Access E E isnot amember of any CUG D does not have Incoming Access X890 304751 Issue 1 D 4 Rev 0 D 2 3 Permissions For a DTE Which is Not a CUG Member DTE E can make calls to A Eis not a member of any CUG A has Incoming Access C E isnot a member of any CUG C has Incoming Access DTE E CANNOT make calls to B Eis not a member of any CUG B does not have Incoming Access D E isnot a member of any CUG D does not have ncoming Access DTE E can receive calls from D E isnot a member of any CUG D has Outgoing Access DTE E CANNOT receive calls from A Eis not a member of any CUG A does not have Outgoing Access B E isnot a member of any CUG B does not have Outgoing Access C E isnot a member of any CUG C does not have Outgoing Access X890 304751 Issue 1 D 5 Rev 0 X890 304751 Issue 1 D 6 Rev 0 Appendix E ACS Support This appendix provides a brief outline of the functionality of the ACS Access Control Server and how it is supported by Xpress E 1 The ACS The ACS provides a network security service for an Xpress network It does this by intercepting user calls and presenting the user with a menu driven service election session In this way users can be prevented from connecting directly to hosts etc Once the user has been connected to the ACS and identified him
156. including port control parameters PVC details CUGs to which this port belongs called and calling address translation tables applicable to X 25 ports only Module configuration report including module type dump reload on failure flags module version numbers buffer pool sizes recovery and danger levels Application configuration report including list of installed applications list of management addresses Logical port allocation report this lists all the configured logical ports and their physical locations Printer port configuration report this lists the current printer port configuration details line speed parity setting bits per character etc To get to the printer port configuration screen from the Main Menu select the Configuration Port Configuration Local Printer Configuration screen The reports are output on either the local node or central network printer depending on where printing is being directed to X890 304751 Issue 1 5 14 Rev 0 5 7 Events This section describes how event information is available to the user of an Xpress PSE Event information is part of the general monitoring information available to help the node operator in administration performance monitoring and fault diagnosis Other such services are Statistics Status Billing and Charging information Events are automatically generated operator notifications about some change in the state of the system Events may be handle
157. inks is madein any of the status display screens This is because X 25 X 75 trunk circuits are carried transparently over frame relay links and the fact of a frame relay circuit being down will always be reflected in the status of the links being carried over it 5 11 1 Display Node Status This screen displays the type and state of each module in the node Under normal circumstances a module should be in the Operational state Two types of module are used within the node the Utility Module UM running the node management system and the X 25 X 75 Interface Module XIM running the X 25 X 75 switching software Each module within the system will bein one of the following states Unknown The management system cannot communicate with the module not powered up total failure etc Idle The module has just completed its powerup diagnostics and is now ready to accept a load image from disk Software error The module has failed due to a software crash Dumping The module has failed due to a software error A post mortem dump file is being created for the module on disk Loading The module is in the process of being loaded with software from disk Operational The module has been loaded with software from disk configured and is now running This is the normal state a module should be in UPM hardware failure ACM hardware failure The module has a hardware failure X890 304751 Issue 1 5 25 Rev 0 Call Operator vvnen In tnis
158. internal routing for example PVC setup Call re establishment central printing node manager etc This DNIC is now configurable The menu Configure Node Configure Edit node configuration Internetworking DNIC allows entry of the new DNIC This menu item is also used to configure the X 75 DNIC in this and previous versions Care should be taken if X 75 is used or in an earlier version e g 8 5 the Internetworking DNIC had been accidentally configured The node will no longer respond to traditional X 121 numbers like 11000019000 X890 304751 Issue 1 O 1 Rev 0 X890 304751 Issue 1 O 2 Rev 0 Appendix P Network User Identification P 1 Introduction Support of the X 25 facility Network User Identification NU A has been expanded with Version 9 The NUI field can now be checked on a per port basis and if acceptable a suitable calling NUA Network User Address substituted to identify the user and the call allowed to proceed Typically 100 NUI s per port are configurable Configuration is similar tothe existing ICAT and OCAT tables X890 304751 Issue 1 P 1 Rev 0 P 2 Parameters to be Configured This feature is configured on a per port basis using the menu Routing specification Network User Identification Table entries are for the NUI and the substitute NUA A NULL NUA will disable the corresponding NUI To enable the feature a corresponding entry must be made for that port using the menu Configuration Port Configuration X 25 X
159. ions DTE B can make calls to C BandC arebothin CUG1 B allows outgoing calls within CUG 1 C allows incoming calls within CUG 1 DTE B CANNOT make calls to A AandB arebothin CUG1 B allows outgoing calls within CUG 1 A has Incoming Calls Barred within CUG 1 D BandD arenot in the same CUG B does not have Outgoing Access E E isnot amember of any CUG B does not have Outgoing Access DTE B can receive calls from A AandB arebothin CUG1 B allows outgoing calls within CUG 1 A allows incoming calls within CUG 1 DTE B CANNOT receive calls from C BandC arebothin CUG 1 C has Outgoing Calls Barred within CUG 1 D BandD arenot in the same CUG D has Outgoing Access B does not have Incoming Access E E isnot amember of any CUG B does not have Outgoing Access X890 304751 Issue 1 D 3 Rev 0 D 2 2 CUG 2 Permissions DTE D can make calls to A AandD arenot in the same CUG D has Outgoing Access A has Incoming Access C Cand D are both in CUG 2 D allows outgoing calls within CUG 2 C allows incoming calls within CUG 2 E E isnot amember of any CUG D has Outgoing Access DTE D CANNOT make calls to B BandD arenot in the same CUG D has Outgoing Access B does not have Incoming Access DTE D can receive calls from C CandD areboth in CUG 2 C allows outgoing calls within CUG 2 D allows incoming calls within CUG 2 DTE D CANNOT receive calls from A AandD arenot
160. irectly with the NMC and not indirectly via the PSE Node Manager The NMS service is accessed with the AF_X25 SOCK _NMS socket J 2 3 Network Connectionless Service A future version of Xpress will allow applications to exchange connectionless messages across the Xpress network using the Xpress network connectionless service This service is accessed using the AF_X25 SOCK _DGRAM type of socket J 2 4 Node Connectionless Service A future version of Xpress will allow applications within the same Xpress PSE to exchange connectionless messages using the Xpress internal node connectionless services Xpress transfers these messages using a fast mechanism which does not guarantee reliable delivery This service is accessed using the AF_BUS SOCK _RAW type of socket X890 304751 Issue 1 J 2 Rev 0 J 3 Hardware Architecture ACM is the generic name for boards which directly connect toUPMs Some types of ACM do not have any on board processors i e they are completely controlled by the UPM processor Other types have one or more on board processors and are called intelligent ACMs The ACM processor s belong to the 680X0 family of processors J 4 Software Architecture See Figures J 1 to 3 Applications co reside on UPMs with the Xpress software or reside on intelligent ACMs If an intelligent ACM has more than one processor then a separate instance of an application runs on each processor Both types of application in
161. is performed as a call leaves the port X890 304751 Issue 1 4 22 Rev 0 4 6 Closed User Groups CUGs The Closed User Group CUG is a set of optional user facilities subscribed to by a group of ports on a PSE and in its basic form it restricts communications to within a specified group of ports CUG members can also opt for membership which allows them extended access to or from ports belonging to other CUGs or to the open part of the network The number of CUGs on a node is restricted to 99 CUG calls are restricted to PSE ports Across the network as a whole up to 65535 CU Gs can be configured 4 6 1 CUG Membership Criteria Any X 25 port excluding trunks and the virtual DTE can opt to belong toa CUG In addition hunt groups are also allowed to belong to CUGs A port can belong to at most 99 CUGs Each port in a CUG has a particular set of access rights known as a subscription explained in Section 4 6 2 Ports belonging to more than one CUG must specify a preferential CUG at subscription time if they have only the basic CUG subscription i e neither Incoming Access nor Outgoing Access This prevents the port making a call to somewhere outside its CUG s If the ports have either incoming or outgoing access then specifying the preferential CUG is optional During the call setup phase if no CUG is specified the PSE will insert the preferential CUG index If no preferential CUG has been specified or the port does not belong to
162. isables call clearing The rate at which the calls are cleared is determined by the two parameters Call Clearing Interval and Calls Cleared in One CCI see below Note that the use of Clear Thresholds may give rise to an increased number of management events being generated due to the call re establishments taking place e Priority to Clear This parameter governs which priorities of calls are candidates for automatic clearing whenever the Clear Threshold is exceeded The choices are based on priority class relative to the priority class currently selected on the screen The choices are current only calls of the currently selected priority class can get cleared lower only calls belonging to priority classes lower than the one currently selected can get cleared current amp lower Calls belonging either to the selected priority class or lower classes can get cleared Whenever several calls are candidates for clearing the rule is the next call to be cleared is the most recently established of the lowest priority calls currently present Note that the system never clears all of the calls on the trunk or link the last one will always be preserved regardless of its priority and of the configured clear thresholds e Call Clearing Interval CCI e Calls Cleared in one CCI Together these two parameters govern the rate at which the system clears calls At the end of each call clearing interval the port software w
163. issions for the port in the specified CU Gs can then be set by submitting the form press PF 1 4 6 7 Effects of CUG Permissions on Making a Call This subsection explains what happens when an X 25 call is made using CUGs in different circumstances X890 304751 Issue 1 4 25 Rev 0 e If called and calling X 25 ports belong to the same CUG If the called port has Incoming Calls Barred or the calling port has Outgoing Calls Barred then the call will be rejected e If called and calling X 25 ports belong to different CUGs If the called port has no Incoming Access or the calling port has no Outgoing Access then the call will be rejected e If either the called or calling X 25 port does not belong toa CUG Ifthe called port alone belongs to a CUG and has Incoming Access then the call will be allowed If the calling port alone belongs toa CUG and has Outgoing Access the call will be allowed X890 304751 Issue 1 4 26 Rev 0 5 Utilities There are five utility functions which can be selected from the top level utilities menu e Access utilities to set user access attributes and to change passwords e Clock utilities to reset the current date and time e Disk utilities to perform disk and file operations e Print utilities to print off a hard copy of system configuration e Install Delete E xpand application e Dump utilities to remove or analyse system dumps 5 1 Access Utilities The access utilities provide security for the
164. itoring and Control is to be used on one or more inter node trunks within the network e Priority Class Profile This parameter is found on the Congestion Monitoring menu Configuration Port Configuration X 25 X 75 Port Configuration Congestion Monitoring It should be configured on the X 25 X 75 link ports where calls enter the Xpress network It controls the choice of priority class assigned to each incoming call The other parameters on this menu will usually not apply to X 25 X 75 link ports and can be left at their default values Every call entering the Xpress network needs to be allocated a priority level at its entry port in order for the Congestion Monitoring feature to handle it properly Since there is no standard way for a DTE to signal X890 304751 Issue 1 M 6 Rev 0 to tne network WNat a call s priority IS Apress interprets the Throughput Class facility for this purpose A mapping table is used to assign priority levels to calls entering the network based on the value of Throughput Class requested The individual entries in this table are not configurable via the Node M anager instead a choice of two pre configured tables or profiles is provided Select 1 or 2 as appropriate The default is profile 1 The two profiles are as follows Throughput Class bps Profile 1 Profile 2 75 150 300 600 1 4 1200 2400 4800 2 3 9600 19200 3 2 48000 64000 4 1 For Congestion Monitoring to distinguish correctly between the
165. ivided into two pages Note that these options X890 304751 Issue 1 3 18 Rev 0 are relevant to A 25 Tacilities OF A 5d UTHITIES aS appllCable according to the interface configuration They are independent of whether the port is a real physical port or a frame relay virtual physical port First Page e Charging Information X 25 only Setting this parameter to YES causes charging information to be sent to the port being charged for the call normally this is the call originator Default is NO e Reverse Charge Acceptance As for Fast Select call acceptance the PSE will always accept an incoming Reverse Charge call but will only forward it out on a port with Reverse Charge Acceptance set toYES Defaultis NO e Local Charging Prevention X 25 only If this parameter is set to YES then the PSE will ensure that the port is not charged for any calls It will dothis by attempting to make remote ports accept charging If remote ports refuse to be charged then the calls to from the port will be rejected by the PSE The default is NO e Throughput Class Negotiation As for flow control parameter negotiation if this is set to YES the software will allow a call s throughput class to be negotiated to a value different from the default if necessary Default is NO e Default Throughput Class The default throughput class may be set to a range of values from 75 bps to 64000 bps The default is 9600 bps It may be necessary to configure this on
166. k for example to allow interactive terminal users to have preference on it compared to non response critical traffic such as file transfers This is the main purpose for which the feature is intended and the remainder of this chapter assumes this However it can also be used on hunt groups of X 25 or X 75 link ports to supplement the load balancing already provided by the hunt group mechanism In general Congestion Monitoring is beneficial only when it is operating on a trunk or link for which there is at least one alternative route to take up any displaced calls Note that this feature is not related to the Frame Relay congestion monitoring period parameter described in Section 3 4 3 The key elements are e Assignment of Priority Class for a call In order to distinguish between calls at different priorities one of four priority classes is automatically assigned toa call when it is first X890 304751 Issue 1 M 1 Rev 0 establisned ine attacned aevice inaicates tne Friority Class Of tne Call by means of the Throughput Class Facility in the Call Request Packet which the entry port uses to internally map the call into the appropriate priority class A call cannot change from one priority class to another e Measurement of Link Utilisation Level For every trunk or X 25 X 75 link port that is on line Xpress continuously measures the utilisation level monitoring traffic flow in the outward transmit direction This is the cri
167. k starts to become very poor once the error rate exceeds about 8 bits in every 10 The error rate displayed in this case is approximately 35 and in most instances useful values for this parameter will bein the range 5 to 50 e Error monitoring period This defines the length of the time the system leaves between computing each successive value of mean error rate During this interval the system collects raw line error information for use in its X890 304751 Issue 1 3 23 Rev 0 calculation iN oraer to smooth over ISOlated SpiKeS Lo Seconds IS a Suitable minimum value for this parameter Error Monitoring and Control is disabled by leaving this parameter at its default value of 00 00 This also disables the calculating of mean error rate which therefore always appears as 0 while the feature is disabled e Port reinstatement delay If this parameter is set to a non zero value it defines the delay before a port that has closed itself due to errors automatically re opens It is recommended that the port is not allowed to reinstate sooner than the next time the mean error rate is computed in practice this parameter should be set to at least twice the Error Monitoring Period if automatic reinstatement is required The default value is 10 minutes Setting the Port reinstatement delay to zero disables automatic reinstatement altogether In this case operator intervention is required to manually ready the port for use The operator sh
168. ks are normally treated in the same way as physical links See Appendix J for details of applications and application links The PSE end of a physical link is called a physical port The physical ports on a PSE are numbered by bay number always 0 slot number and link number The number of slots per node and links per slot vary between different members of the Xpress range The physical port number range starts at 0 x 0 bay 0 slot x link 0 where x is the slot number of the lowest numbered slot which may contain X 25 cards and extends to 0 y z 1 where y is the number of slots in the system X890 304751 Issue 1 3 1 Rev 0 and z IS tne number OF INKS per stot E g the oOwest and Nignest pnysiCai port numbers on the 8325 which has 5 slots and 6 port cards are 0 2 0 and 0 5 5 respectively 3 1 2 Logical Ports Each physical port has associated with it a logical port number which is a four digit number in the range 0000 to 9999 It is via the logical port number of a physical port that the vast majority of port references within the Xpress operational and management software are made This is because logical ports can easily be swapped between physical ports allowing an alternative physical port to be selected e g in the case of port failure The PSE stores a mapping between logical and physical port assignments which ensures that the logical port s configuration can be automatically applied to any physical port with the correct har
169. l Call collision Duplicate facility or utility expected Non zero address length Non zero facility length Facility or utility expected Maximum no of call redirections or call deflections exceeded BAD CAUSE CODE FROM DTE Non octet aligned Invalid Q bit NUI Problem Table B 4 continued Diagnostic Codes X890 304751 Issue 1 B 5 Rev 0 International call set up problem Unknown calling DNIC TNIC utility mismatch Call identifier utility mismatch Utility parameter negotiation problem Network utility length invalid Non zero utility length M bit violation International problem Remote network problem International protocol problem International link out of order International link busy Transit network facility problem Remote network facility problem International routing problem Temporary routing problem Unknown called DNIC Maintenance action Pf Unknown called DNIC 132 Asynchronous DTE busy 85 133 Asynchronous DTE address invalid Timer expired for interrupt confirmation Timer expired for data packets Timer expired for reject packet DTE operational DTE not operational DTE resource constraint a 176 Miscellaneous X 25 violation 177 X 25 link is down Table B 4 continued Diagnostic Codes X890 304751 Issue 1 B 6 Rev 0 XPRESS ORIGINATED CALL CLEARING Intra node call congestion Inter UPM virtual link is out of order UPM congestion buffer depletion Xpress network congestion Hop Count e
170. least one call of a certain priority class using a link and if at the same time the measured utilisation level on the link exceeds the configured Clear threshold for this class X890 304751 Issue 1 M 2 Rev 0 The system starts to free up bandwidth for this class of traffic by clearing other calls off the link which are usually those of a lower priority although this is configurable Depending on the configured routing the internally cleared calls will get re established over an alternative route transparently to the user The process stops once the link utilisation has decreased below the threshold or if there are no longer any suitable calls left to clear X890 304751 Issue 1 M 3 Rev 0 M 2 Parameters to be Configured This section describes each of the parameters to be considered when configuring Congestion Monitoring an example of the application of the feature is given in section M 3 M 2 1 Configuring the Congested Trunk Port All the parameters described in this section are located on the menu Configuration Port Configuration Trunk Port Configuration Congestion Monitoring and apply on a per port basis Leaving all the parameters at their default values disables the control aspect of the feature from operation although the link utilisation is continuously measured while the port is in service as explained previously These parameters are applicable and should be configured at the trunk ports where congestion monitorin
171. lisation reaches 85 Congestion control is to be introduced on Node 3 T 31 where the congestion is occurring At the other end of the trunk inspection of Node 1 T11 s utilisation gives the traffic level in the other direction i e towards the host This appears rarely to exceed 20 so it is decided that there is no need to enable congestion control at the Node 1 end of the trunk M 3 2 Configuring the Example Network Each of the terminal equipments terminals workstations host application is set up to specify the following throughput classes which map correctly to the required priority classes providing Profile 2 is used workstation 64000bps Priority 1 terminal users via PADs none so port default of 9600 bps applies Priority 2 host initiated printer jobs 2400 bps Priority 3 Priority 4 is not used Therefore each of Node 1 s logical ports 40 50 and 60 has to be configured to use Profile 2 The settings that have been chosen for Node 3 T31 are shown in Figure M 2 X890 304751 Issue 1 M 10 Rev 0 CRAY Node Manager Node 3 Test Node 3 24 Jan 94 18 11 TRUNK PORT CONFIGURATION congestion monitoring port number T0031 Trunk to Node 1 Call clearing interval CCI 01 00 mm ss Calls cleared in one CCI 1 Priority class profile 1 Refuse at Clear at Priority to clear Priority class Priority class 2 Lower Priority class 3 Current amp Lower Priority class 4 Current amp Lower Options Repeat c
172. llowing subsections 5 10 1 Display Port Statistics This menu lets you choose the level of port statistics to be displayed on a real time basis Statistics can be displayed for the packet level frame level or physical level for a specific port Note that for Application Links only packet layer statistics can be obtained Statistics for a given port at all levels of the X 25 X 75 software are automatically reset to zero when any of the three display screens is entered for the first time The exception is if the port is already in the list for fixed interval statistics reporting see Section 5 10 9 In this case the most recent statistics will be displayed The Update option allows you to see the most recent set of statistics for that port You can switch between the physical level frame level and packet level display screens without the statistics being reset You can use Update to refresh the screen display To reset the statistics for the port you have selected either exit the statistics menu totally press PF4 and re enter the display screen or select the Repeat option and specify the same port 5 10 2 Display Physical Level Statistics This screen displays the physical level statistics for a specific port and the current state of its control signals The time since the statistics were last reset is displayed X890 304751 Issue 1 5 20 Rev 0 This screen also displays the status of V 54 test loops and errors detected by
173. lly connected to a port on the card containing that server As explained above the host can be anywhere on the network In the two examples given the host is on Node 1 Logical Port 40 and this logical port could be assigned to any suitable port on the node The same applies to the client connections for each server i e clients do not have to be physically connected to the card on which their server resides although for reasons of efficiency this is likely to be desirable in most applications X890 304751 Issue 1 G 10 Rev 0 These 4 PADs can support up to 31 clients Port 32 Ports 10 13 Ports 20 23 These 4 PADs can support up to 31 clients These 4 PADs can support up to 32 clients Key X 25 link Trunk These 4 PADs can support gt Broadcast data flow up to 32 clients 4 ABS Server on slot 4 Figure G 4 Example of a Multi node Multi server ABS X890 304751 Issue 1 G 11 Rev 0 Figure G 4 shows a 5 node network providing a broadcast facility for up to 126 clients This network has been designed to give an example of an arrangement of ABS servers which will minimise the amount of broadcast data being sent around the network The following notes highlight relevant issues which have resulted in this arrangement e Theclients PADs are grouped around four cards The cards to which each group of PADs is physically connected are used to run the server for each group of four This means that as far as p
174. longs 4 6 5 Specify CUG Membership for a Logical Port This menu enables you to set up CUG membership for X 25 ports and define their access levels 1 Select the Specify CUG Subscription for Logical Port screen Select the required option to create edit or remove a subscription 2 Tocreate a new subscription select the Create new subscription for logical port screen Enter a configured port s number or hunt group address excluding trunk ports and the virtual DTE Enter the local CUG s to which the port belongs 3 By default intra CUG permissions are set to Two way To change this use the Change access permissions within CUG for logical port Screen 4 Extra CUG permissions i e the per port access levels can be changed if required If the extra CUG permission is set to No external access the default value you must specify a preferential CUG index 5 To edit or delete a subscription follow the prompts to change the local CU G s or the CUG permissions as required 4 6 6 Change CUG Subscription 1 Select the Change access permissions within a CUG for a logical port Screen 2 Enter the port address the current access permission remember that the default is Two way access and the required permission 3 Enter the CUGs to which the port belongs in the current configuration if you ve forgotten then typing will display these and you can delete the ones which you don t want to have the new permission The CUG perm
175. low control The core frame relay frame is very similar toa LAP D information frame and is shown in Figure A 7 The address field is similar to a 2 byte LAP D address with the SAPI TE field re defined to hold a 10 bit Data Link Control Identifier DLCI The DLCI is used to uniquely identify each end to end procedural Level 2 connection cr is the LAP command response bit X890 304751 Issue 1 A 23 Rev 0 DDDDDDc r0 DDDDfbdt Figure A 7 The Frame Relay Frame f and b are the forward and backward explicit congestion notification bits FECN and BECN which are used by the network to indicate a congestion condition in the direction of transmission to and receipt from the network respectively These bits are used to notify destination and source flow controlled DTEs respectively to reduce the traffic they are sending across the interface according to rules defined in the frame relay implementation standard d is the discard eligibility bit Data with this bit set is discarded by the network in preference to non discard eligible data in the case of congestion In brief forum frame relay networks use this single frame type to carry higher layer protocols transparently end to end between end systems over pre defined frame relay PVCs identified by DLCIs which are locally unique and are mapped end to end by the network Flow control is also performed end to end with the network responsible for notifying congestion conditio
176. maining on the disk X890 304751 Issue 1 5 6 Rev 0 The floppy disks supplied with the PSE contain the following files v7boot initial bootstrap program of 8425 8525 v7xboot initial bootstrap program of 8325 data system data files L system load files During configuration the PSE will create the following files config system configuration files empty temporary work file Following a module failure a dump file may be generated see Section 5 4 core b s dump file for module bay slot s At any stage a different disk can be selected by using the Specify another disk option Next page and Previous page allow thelist to be scanned in fixed amounts 5 3 4 File Copy The file copy command allows files to be copied between local and or remote disks The syntax of this command is copy from_file to_file The files must be specified as drive filename although when specifying the target file just the drive is sufficient the file name being defaulted to the source file name A file cannot be write protected You are not allowed to copy a file onto itself The syntax for a file is Dnnn filename where D is mandatory and should be either a or b for drive a or drive b where nnn is optional and is the numeric node number e g 12 345 where filename iS up to 14 characters long and consists of the set of characters a to z AtoZ 0to9 and Pattern matching can be applied to the from filename only to match any sequen
177. mechanism is described in Appendix E Any connected device may use this facility if it is capable of doing so Call Deflection Referral This non standard facility allows an unsuccessfully deflected call e g cleared by device deflected to to be referred back to the deflecting device usually the ACS which can then try another deflection if possible This mechanism is described in Appendix E X890 304751 Issue 1 A 13 Rev 0 18 Call Redirection This is an X 25 1984 88 User Facility The redirection address Alternate Network Address can be any X 121 number of between one and 15 digits in length This address may be modified by the Address Translation functions if required 19 Network Data Integrity This non standard facility allows the PSE to recover data which may have been lost within the network after a call has been re routed or internally reset Calls using D bits are automatically protected by the network data integrity facility Note use of Network Data Integrity may reduce the maximum packet throughput of the port and may lead to packet buffer shortage on UPM cards with only 1 Mbyte of memory Network Data Integrity can only be provided for calls that traverse nodes all of which are running version 5 or later software RPOAs RPOA Subscription is a 1988 facility which allows the operator to configure the PSE toinsert at most one RPOA Selection into an outgoing call request Xpress supports both the Basic and Extende
178. n Code CNIC This identifies the clearing network If Xpress clears an established call then it inserts its Internetworking DNIC into the clear request Xpress will transfer a CNIC over X 75 1984 interfaces and discard it at X 75 1980 interfaces Xpress provides a configurable per X 75 port option to suppress CNIC insertion d Traffic Class Indication Xpress transfers this transparently over X 75 interfaces e Transit Delay Indication Xpress transfers this transparently over X 75 interfaces f Unrecognised Network Utilities Xpress transparently transfers unrecognised Utilities g Non X 75 Network Utilities These are preceded by a special marker of value zero Xpress transfers these non standard Utilities transparently X 25 User Facilities These are X 25 facilities which have been moved from the User Facilities field into the Network Utilities field before a packet is transferred across X890 304751 Issue 1 A 18 Rev 0 an A o interface I Nese Tacilities are Moved back to tne User Facilities field before the packet is transferred across an X 25 interface Note that the options on the User Facilities screen apply to the outgoing X 25 facilities or the equivalent incoming X 75 utilities Note also that Xpress transparently transfers the User Facilities field across an X 75 interface apart from checking that the facilities within it have correct format syntax Xpress supports the following Network Utilities which are m
179. nal statistics will not be available Note that if this sreen is selected for a logical port that maps toa frame relay virtual physical port then the statistics displayed will be for the physical port over which the virtual physical port is being multiplexed 5 10 5 Frame Level Port Statistics This screen displays the frame level statistics for a specific port The time since the statistics were last reset is also displayed X890 304751 Issue 1 5 21 Rev 0 If the module supporting this port is not operational statistics will not be available 5 10 6 Packet Level Port Statistics This screen displays the packet level statistics for a specific port The time since the statistics were last reset is also displayed If the module supporting this port is not operational statistics will not be available 5 10 7 Modify Report This menu lets you select which statistics report you want to change Three reports are available Link statistics Module statistics for Cray engineers use only INCS statistics for Cray engineers use only All reports are printed at the same time this time being governed by the statistics reporting interval 5 10 8 Link Statistics Report This screen lets you specify which logical ports are to be included in the link statistics report If no ports are included the report is not produced Ports may be added to or deleted from the report The wild card character provides a quick means of adding deleti
180. nalysis Table must be configured at each node to direct calls to this X 25 gateway port The X 25 gateway node must have a PDN Gateway Table set up Incoming and Outgoing Call Address Translation Tables may be required for the X 25 gateway port 3 9 7 X 75 Gateway To configure a port as an X 75 gateway the following options must be selected as specified These are in addition to the port configuration as defined in Section 3 4 The following options must be selected on the configuration Port configuration X 25 X 75 port configuration Data Link level Screen Protocol Option 1 for CCITT handling X 75 support Select correct X 75 support The following options must be selected on the configuration Port configuration X 25 X 75 port configuration Network level Screen Profile dentifier One of the CCITT X 75 profiles The following options must be selected on the configuration Port configuration X 25 X 75 port configuration User facilities screen DNIC of Local RPOA DNIC of attached network TNIC Suppression As required CNIC Suppression As required PDN Gateway Table entries or Address Analysis Table must be configured at each node to direct calls to this X 75 gateway port The X 75 gateway node must have a PDN Gateway Table set up Incoming and Outgoing Call Address Translation Tables may be required for the X 75 gateway port Thelnternetworking DNIC must be configured on the X 75 gateway node This is selected
181. nfiguration change completed successfully will be displayed Press PF3 to return tothe X 25 Port Configuration menu To keep things simple assume that the application requires no changes to the user facilities screen although you may care to look at it and to try switching flow control parameter negotiation to YES Use the change state of port screen to bring the application link Online Enter the Configuration Node Configuration detailed link status display screen and verify that the port is Online and up Now try the call X890 304751 Issue 1 3 36 Rev 0 OnPAD 1 sterminal type CON 1100001700301 RETURN This should cause the terminal to display the application s MMI Select an application command to ensure that traffic flows in both directions Toclear the call type CTRL P CLR PAD 1 s terminal should display CLR CONF X890 304751 Issue 1 3 37 Rev 0 3 7 Curing Problems This section attempts to describe some of the problems that may for one reason or another occur with the above procedure 3 7 1 Node Does Not Power Up It is assumed that the initial RETURN to wake up the manager terminal failed to do so 1 The most likely cause is that the manager terminal is incorrectly connected or set up Ensure that the cable connecting the terminal to the manager port is secure Set the terminal to 8 bits No parity and 1200 2400 4800 or 9600 baud 2 Thesystem disks could be missing inserted the wro
182. ng all configured ports from the report list The frequency at which this report is printed is governed by the statistics reporting interval For each logical port included in the report the following details are printed physical port address bay slot link packet level statistics frame level statistics physical level statistics 5 10 9 Module Statistics This screen lets you specify which modules are to be included in the module statistics report If no modules are included the report is not X890 304751 Issue 1 5 22 Rev 0 proauced Moauies May be aaaged to or ageieted trom tne report Ne wild card character provides a quick means of adding deleting all modules in the report list Modules are identified by their slot number The frequency at which this report is printed is governed by the statistics reporting interval For each module configured into the report the following details are printed module type e g XIM1 module state e g Operational memory buffer pool usage processor idle times These statistics are for Cray use only 5 10 10 Intra node Communications Subsystem INCS Statistics This screen lets you specify which modules are to be included in the INCS statistics report If no modules are included the report is not produced Modules may be added to or deleted from the report The wild card character provides a quick means of adding deleting all modules Modules are identified according to t
183. ng way round write protected or faulty 3 Themanager hardware could have failed its power up diagnostics 4 TheUPM associated with the UM is the wrong type i e has only 1 Megabyte of RAM 3 7 2 XIMs Not Loaded If the node status display screen at step 7 above does not show two XIM1s in Operational state the loading procedure has not completed successfully 1 TheXIM load file on the disk in drive A could be missing or faulty Try a different disk 2 TheXIM may havea hardware error in which case the state will be ACM h w failure Try a different XIM 3 TheUPM may havea hardware error in which case the state will be Unknown or UPM h w failure Try a different UPM This could also be caused by a problem with the backplane or even because the boards are not plugged in properly 4 Any other state generally indicates a XIM UPM software failure A persistent error will result in the state settling at Software Error or Call Operator Contact your supplier X890 304751 Issue 1 3 38 Rev 0 3 7 3 Errors During Configuration The menu system will not allow you to apply illegal configurations to ports e g it will not let you assign more than 256 LCNs toa XIM UPM1 or to set parameters to bad values Level 2 window size 537 etc Therefore if a screen is not confirmed when you pressed PF 1 check for typing errors Never remove the disks during configuration as the configuration changes will not be recorded because there i
184. ngle physical port The physical ports on a single card can be internally disconnected from the hardware and the traffic redirected to a single physical port on the same card which provides a single interface to the frame relay network The traffic from each of these FR virtual physical ports is then carried over the frame relay network identified by a unique frame relay Data Link Connection ID DLCI This identification of the individual virtual physical ports via DLCI allows them to be mapped through the frame relay network to different remote ports on that network E g anode may havea single frame relay network interface carrying two virtual ports which are mapped to two hosts one in London and one in New York The frame relay network does not itself understand X 25 but carries the X 25 level 2 and 3 protocols transparently between the Xpress port and the X 25 device inside frame relay frames This technique is Known as encapsulation and is explained in Section A 3 Standard Xpress Ports Xpress Ports multiplexed over Frame Relay Each port s traffic runs directly over its own physical link and is Each port s traffic is encapsulated identified by the physical link within a frame relay data link number connection on physical link 0 and is identified by a DLCI Figure 3 1 Xpress Virtual Physical Ports X890 304751 Issue 1 3 3 Rev 0 For example see Figure 3 1 which shows how the six ports on an 8325 SAC card in slot 2 s
185. nk Ports M 2 3 Configuring Trunk Ports on Secondary Routes M 2 4 Trunks toPreVersion 9 Nodes Using Congestion Monitoring and Control M 3 1 Description of the Example Network M 3 2 Configuring the Example Network M 3 3 Congestion Monitoring Takes Effect on the Trunk Summary Points Error Monitoring and Control ntroduction Parameters to be Configured Configurable DNIC Network User Identification ntroduction Parameters to be Configured Figures Xpress Virtual Physical Ports Example Frame Relay Trunk Configuration ISO 7 Layer Model for Open Systems Interconnection HDLC Frame Structure Level 3 Packet Structure Call Procedure Using an SVC X 25 Switching Frame Relaying The Frame Relay Frame X 25 Encapsulation Trunk Protocol Encapsulation Example of CUG Groupings X890 304751 Issue 1 0 12 M 4 M 4 M 6 M 8 M 8 M 9 M 9 M 10 M 12 M 14 N 1 N 1 N 2 O 1 P 1 P 1 P 2 3 3 4 10 A 3 A 4 A 7 A 23 A 23 A 24 A 26 A 27 D 1 Rev 0 C L c xampie ACS Network E E 2 Reselection PAD Message F ormat E 5 F 1 Modem Test Loops F 2 G 1 Example of a Single Node Single Server ABS Configuration G 6 G 2 Example of a Linear Multiple Server G 9 G 3 Example of a Hierarchical Multiple Server G 10 G 4 Example of a Multi node Multi server ABS G 11 G 5 Example of Multi Service ABS G 15 J 1 UPM Co Resident Application J 4 J 2 ACM Application on a Card with One ACM Processor J 5 J 3 ACM Application
186. ns to the end systems which are then expected to take action according to well defined rules to reduce traffic flow If this does not take place or does not do so fast enough the network will discard data The Local Management Interface LMI is used by this PVC based system to allow the network and the DTE toidentify PVC assignment failure reliability etc The LMI is a simple message based system which runs on DLCI 0 It is basically a highly cut down version of the full ISDN control plane signalling system used by full SVC based frame relay systems X890 304751 Issue 1 A 24 Rev 0 A 3 3 Series 8000 PSEs and Frame Relay Xpress follows all the rules to do with the core and LMI aspects of frame relay DTE support as laid down by the frame relay forum i e the Xpress implementation is compatible with the mandatory requirements of the standards identified in the Frame Relay Forum Technical Committee document Frame Relay X 25 Interworking Implementation Agreement FRFTC 92 15 specifically data transfer ANSI T1 618 congestion control procedures ANSI T1 618 Annex A and LMI procedures ANSI T1 617 Annex D Xpress supports explicit congestion avoidance and control via the BECN bit and a dynamic procedural Level 2 window size The slow start mechanism is also supported Xpress does not use the discard eligible bit mplicit congestion notification is supported by means of the procedural Level 2 LAP B noticing frame loss Xpress handles thi
187. nslation 4 5 3 Incoming Called Calling Address Translation I CAT 4 5 4 Outgoing Called Calling Address Translation OCAT 4 6 Closed User Groups CUGs 4 6 1 CUG Membership Criteria 4 6 2 Access Levels within CUGs 4 6 3 Setting up CUGs 4 6 4 Configuration of Local to Global Indices 4 6 5 Specifying CUG Membership for Logical Port 4 6 6 Change CUG Subscription 4 6 7 Effects of CUG Permissions on making a Call 5 Utilities 5 1 Access Utilities 5 1 1 Change User Password 5 1 2 Type Specification 5 1 3 User Access Menu 5 1 4 Initial Users 5 2 Clock Utilities 5 2 1 Change Date 5 2 2 ChangeTime 5 3 Disk Utilities 5 3 1 Format Disk 5 3 2 Copy Disk 5 3 3 List File Directory 5 3 4 File Copy 5 3 5 Remove File 5 3 6 MoveFile X890 304751 Issue 1 0 7 Rev 0 5 4 5 5 5 6 5 7 5 8 5 9 5 10 5 11 aTe i Verify DISK 5 3 8 Automatic Disk Verification Dump Utilities 5 4 1 Delete Dump File 5 4 2 Print Dump File Install Delete E xpand Applications 5 5 1 Display 5 5 2 Installation 5 5 3 Deletion 5 5 4 Expand 5 5 5 Background Information 5 5 6 Application Specific Files Print Utilities Events 5 7 1 Alarms and Warnings Charging Billing 5 9 1 How Billing Works 5 9 2 Configuration Statistics 5 10 1 Display Port Statistics 5 10 2 Display Physical Level Statistics 5 10 3 Frame Relay Core Level Statistics 5 10 4 Frame Relay LMI Statistics 5 10 5 Frame Level Port Statistics 5 10 6 Packet Level Port Statist
188. numbers lt failure code gt The error code is repeated several times before an F for failure is displayed on the panel A common cause of diagnostics failure is to boot up the system without any disks in the drive Your supplier will be able to advise in case of diagnostic failure Stage 2 The manager code is loaded After hardware diagnostics are completed the UM will display b for booting This lasts a few seconds and then changes toL The manager code is now being loaded While this is taking place the other boards will only display a flashing dot Stage 3 The rest of the system is booted up Once the manager is loaded up the UM displays initialisation The other slots in the bay are polled in turn to determine the topology of the bay Each XIM indicates P to show that it has been polled The UM then loads the XIMs UPM3 based XIMs are loaded first followed by UPM 1 and UPM2 based XI Ms in parallel and then by application cards The cards will display L for loading Stage 4 Ready to go The LEDs on the cards will successively display o for operational and very soon afterwards I for initialisation andr for running The UM display shows the same transition but with a slight delay after the other boards The system is operational when the UM displays r 2 1 3 If Boot up Fails If you can log on to the system then the Node M anager is operational and it is possible to establish the reason for failure of any of the othe
189. o copy the node manager to the remote node with a different filename You can then issue a move command to rename the file in one operation The following file operations will achieve this Login at node 1 Copy from file b nmU03Um L to file b30 NEWnmU03Um L Login at node 30 Move file b NEWnmU03Um L _ to file b nmU03Um L 10 Wearenow at the stage where the main files have been transferred If you do not have a TGate application installed then skip to step 15 X890 304751 Issue 1 L 6 Rev 0 11 12 13 14 15 16 17 The application consists of a load file and two distribution files For the TGate the load file name is TgateU 03X L the distribution files are applic dist for application data and novid dist for novram data Make sure that the appropriate application distribution disk is present in node 1 drive b Now issue the following file copy operations to copy the distribution disk files to the remote node Copy from file b Tan Lt to file b30 The pattern matching from filename matches T gateU 03X L applic dist and novid dist Logout of node 1 and login to node 30 The application distribution disk is effectively present in drive b now since the files were just copied over from node 1 Hence all that remains to be done is a Utilities Install Delete Expand Application Screen command Use this command to reinstall the application When it prompts with Please insert DISTRIBUTION disk in drive b and press
190. ocal nodes drive a to remote node 123 drive a Copy a123 help data b this will copy the help file on the remote node 123 to the local node drive b Copy a g a123 this will copy all the files exceot the configuration files from the local node drive a to the remote node 123 drive a 5 3 5 Remove File This screen lets you delete files from a specified disk The files are specified as drive filename Several files may be deleted together using a pattern matching character as explained in Section 5 3 4 Once deleted files cannot be recovered X890 304751 Issue 1 5 8 Rev 0 Filenames may be up to 14 characters long and made from the character seta toz AtoZ 0to9 and_ A pattern matching character may be used to match any sequence of characters e g Remove b core will remove all module dump files from drive b Remove a will remove all files from drive a 5 3 6 Move File This screen lets you move rename a single file The file must be on the same disk Remote node numbers are not valid for this command Filenames may be up to 14 characters long and made from the character set a toz A to Z 0to9 and_ Pattern matching characters are not allowed Some example commands are Move b core 0 1 b core save 1 this will rename the slot 1 dump file Move a x2511 config a x2511 cfg this will rename the X 25 layer 1 config 5 3 7 Verify Disk This screen allows a disk to be verified immediately A single drive letter should be ente
191. oftware will insert a number of extra flags between frames to reduce the rate Note that transmit flag insertion has no effect at speeds below 19200 bps e Enable Test Loopback This parameter may be used to enable V 54 modem test loops at this port LOCAL places the local modem into loopback REMOTE places the remote modem into loopback The default is loopback disabled See Appendix F for details of V 54 test loops e Generate Test Pattern If this parameter is enabled the PSE will generate a continuous test pattern to the attached modem If used in conjunction with a modem X890 304751 Issue 1 3 11 Rev 0 test 100p NE FSE WII Monitor the ioopea back Gata Tor errors see Appendix F for details of the test pattern generator e Monitor Test Indicator Signal If this parameter is enabled the PSE will monitor the Test Indicator signal generated by a V 54 modem An event will be raised whenever the state of the signal changes By default the signal is not monitored 3 4 2 Frame Relay Core Level Parameters These parameters are applicable only to logical ports which are mapped to FR physical ports i e those directly interfacing to a frame relay network over which multiple FR virtual physical ports are multiplexed The parameters control the operation of the port with respect to the frame relay core level 2 and Local Management Interface functions Note that if the frame relay core level configuration screen is selected for a logical
192. ommand Keep field values and repeat command for help PF1 submit form PF3 previous menu PF4 main menu Select field name Alarms Warnings Current 0 0 Cleared Figure M 2 Congestion Monitoring Configuration for Node 3 T31 It can be seen that the refuse threshold for priority 2 calls has been set lower 73 than any clear thresholds that can cause these calls to be cleared in this case the 75 priority 1 clear threshold This prevents a looping condition from occurring in which a priority 2 call having just been internally cleared from its primary route trunk port instantly re establishes on the same port only to be cleared again a minute later and so on The Priority Class Profile parameter is irrelevant on the trunk T31 and so it has not been altered from its default value In this network Auto Rerouting is required to take place every 3 minutes to provide regular opportunities for previously displaced calls to be restored to the primary route Therefore the Auto Reroute Interval parameter is set to 3 minutes on each of the trunk ports comprising the secondary route Node 3 T 32 Node 2 T 22 Node 2 T 21 and Node 1 T 12 X890 304751 Issue 1 M 11 Rev 0 M 3 3 Congestion Monitoring Takes Effect on the Trunk Once the trunk Node 1 T11 Node 3 T31 is put on line all 7 terminals Priority 2 and 1 printer job Priority 3 establish calls and utilisation of the trunk increases A typical sequence of even
193. on a Card with Two ACM Processors J 5 K 1 Example of Dial up Link and Trunk Usage K 1 M 1 Example Small Network with Trunk Congestion M 9 Occurring M 2 Congestion Monitoring Configuration for Node 3 T31 M 11 M 3 Utilisation Graph for Node 3 T31 M 13 Tables 3 1 8325 E xample Native Applications 3 7 3 2 8425 8525 Native Applications 3 8 A 1 X 2 Subscription User Facilities A 9 A 2 X 2 1988 Per call User Facilities A 11 B 1 Clearing Cause Codes B 1 B 2 Resetting Cause Codes B 2 B 3 Restarting Cause Codes B 3 B 4 Diagnostic Codes B 4 C 1 X 25 and X 75 Billing Information Record C 2 C 2 X 75 only Part of the Billing Information Record C 4 D 1 CUG Membership D 1 D 2 CUG Call Permissions D 2 X890 304751 Issue 1 0 13 Rev 0 M 1 Profile of Calls Present on Node 3 T31 M 13 X890 304751 Issue 1 0 14 Rev 0 X890 304751 Issue 1 0 15 Rev 0 1 Introduction The Cray Xpress Packet Switching Exchanges PSEs are high performance packet switches capable of processing up to 4096 simultaneous calls between up to 96 X 25 X 75 ports depending on model and configuration Networks of up to 1000 PSE nodes may be configured The ports may be individually configured to offer the facilities described by CCITT Recommendations X 25 1980 X 25 1984 X 25 1988 X 75 1980 or X 75 1984 In addition it is possible to carry any one of these protocols together with the Xpress inter node trunk protocol over a frame relay network in accordance with CCITT
194. opped X890 304751 Issue 1 K 3 Rev 0 RI Pin 22 circuit 125 Ring Indicator Pulsed by the DCE to indicate an incoming call to the X 25 X 75 trunk port The port responds to RI by raising DTR at the end of the first ring i e on a negative transition of RI Note that RI must stay low for at least 0 75 seconds This mechanism is optional as the DCE may signal an incoming call by raising DCD see below Important Note V 54 modem test loops are incompatible with dial up ports that wish to use RI toindicate an incoming call This is because the V 54 Test Indicator signal normally pin 25 is detected on pin 22 by use of a special V 54 cable V 54 test loops may be used in conjunction with DCEs that can indicate an incoming call by raising DCD and seeing DTR go high in response DCD Pin 8 circuit 109 Data Carrier Detect The DCE raises DCD to indicate An incoming call an optional alternative to RI Successful establishment of an outgoing call i e raised in response to DTR after training sequence is completed DCD is dropped by the DCE to indicate circuit failure or call cleared by the remote end Note that the dial up software does not rely on DCD going high to detect a link establishment this it does by detecting the level 2 SABM UA and level 3 Restart Confirm exchange on the link However the software does act on DCD going low treating this as a link failure and will drop DTR in response This software will work corre
195. ork X890 304751 Issue 1 3 12 Rev 0 WItN a FUT otatus Enquiry message INIS Will Cause the network to respond with a message containing details of all currently configured DLCls Xpress uses this information to decide whether or not a particular DLCI is configured and active as far as the frame relay network is concerned The valid range is 1 to 255 heartbeat polls with a default of 6 e Error Threshold N 392 This is the maximum number of LMI reliability i e lost frame or protocol i e bad frame errors which will be accepted by the network or the Xpress port within a sliding Monitored Events count as defined by N 393 See below before the link is declared inactive The valid range for this parameter is 1 to 10 errors per sliding monitored events window with a default of 3 e Monitored Events Count N393 From the network perspective a Monitored Event is the receipt of a Status Enquiry message from Xpress port i e a heartbeat poll or full status enquiry or the expiry of T392 see below From the Xpress port s perspective a monitored event is the transmission of a status enquiry message If more than N392 errors are encountered by Xpress or the network during the sliding window of monitored events defined by N393 then the link is declared inactive by Xpress or the network appropriately Once the link is declared inactive then N393 successful status poll exchanges must be made before the link is again declared active
196. ossible data which is duplicated need not be sent between cards and thus congest the bus For example all the clients on node 3 call 1100 003 9999 005 to access the correct server e Within nodes 3 and 5 one server broadcasts to the other this is to cut down on the amount of data transiting the trunks between nodes 3 and 2 and nodes 5 4 and 2 servers 1 and 2 in node 4 could both be broadcast to directly by server 9 in node 2 but this would result in two copies of the data transiting the trunk e There may be other users on the network who may wish to receive broadcast data occasionally As long as there is a server on the network with spare capacity this is fine For example there is a casual client PC connected to a port on node 4 who wishes to connect to the broadcast service This is no problem as even assuming the four servers on nodes 3 and 5 are completely busy server 9 on node 2 has plenty of spare capacity and the user on node 4 has simply to call 1100 002 9099 005 to connect to server 9 e AAT entries Node 1 Match Address Internal Address 1100 002 9099 995 1100 001 0032 Maps node 2 server 9 s host call to the host computer s port Note that outgoing called address translation may be needed on port 32 Node 2 Match Address Internal Address 1100 002 9099 995 1100001 1100 003 9079 995 1100 002 9099 005 1100 005 9019 995 1100 002 9099 005 Map server 9 s host call to node 1 and nodes 3 and 5 s host calls to server
197. ould put the port fully out of service and then back on line Changing a port s state can be done by any of the means described in 3 4 8 5 Note The fact that the system reinstates a port automatically should not be taken toimply that the bad error rate condition has passed the reinstatement is triggered simply be the configured delay time expiring The act of reinstating a port does not include the restoral of previously displaced calls back toit On trunk ports this is achieved by configuring Auto Rerouting on the secondary tertiary trunk port 3 4 8 Configuration Procedure 1 Set the required application type for the module on which ports are to be configured according to whether frame relay virtual physical ports are required See Section 3 3 2 Create logical ports for the required physical and virtual physical ports See Section 3 2 3 Access the X 25 X 75 Application port configuration screens via the Configuration Port Configuration X 25 X 75 Application Port Configuration menu X890 304751 Issue 1 3 24 Rev 0 4 The Physical Core Frame Relay Data Link Network and User Facility screens may be set up in any order but the following sequence is suggested i Set the physical layer of all X 25 X 75 ports and if present the frame relay physical port ii Set up the Core Frame Relay layer of the frame relay physical port if present iii Set up the Data Link layer of all X 25 X 75 ports including the DLCI an
198. oups and each hunt group can have a maximum of 16 members 3 8 1 Hunt Group Addressing Hunt groups are identified and addressed by a logical port number in much the same manner as ordinary ports However hunt groups use a special range of logical port numbers from 8000 to 8999 This range is not available to ordinary logical ports A hunt group can contain either X 25 ports or trunk ports but not both Hunt groups numbered from 8000 to 8999 contain only X 25 ports Hunt groups numbered T8000 to T8999 contain only trunk ports When configuring a hunt group the individual ports in the group are identified by their logical port number The ports all retain their normal address so calls may still be made to specific ports 3 8 2 Call Distribution within a Hunt Group By default within a hunt group calls are distributed in a simple round robin fashion each subsequent call being forwarded to the next member port in turn However to cater for member ports of different speeds the default action can be modified by specifying a weighting factor for each port Thus a port with a weighting factor of 2 will be forwarded twice as many calls as a port with the default weighting factor of 1 Weighting factors can be specified in a range from 1 to 10 For example X890 304751 Issue 1 3 42 Rev 0 Hunt Group 8000 port number weighting port number weighting 0020 0021 0022 0030 A Rr UN NNNVNN SN 3 8 3 Trunk Groups Hunt group
199. pping for the Xpress network address The first four digits are the Xpress internal DNIC 1100 followed by at least seven more digits in the tagged expression The first three digits of the tagged expression gives the node number the next four digits of the tagged expression gives the port number The Xpress Network Addressing scheme must be allowed for Call Re establishment and Remote Printing to work When a call is cleared due to a network failure the Xpress address of the two ends of the call is used to re establish the call j Any remaining address will be mapped to NULL and will thus be cleared This provides a measure of security but the Xpress network addressing scheme i must be allowed Note that there is great scope for looping calls when the Address Analysis Tables of each PSE are not correctly matched Also beware of Secondary or Tertiary routes to a destination node 4 5 2 Address Translation The main use of address translation is to translate foreign addresses to network addresses as defined by the Address Analysis table of each PSE and vice versa Calls forwarded to a host computer can have their addresses modified to whatever format is demanded by the host computer X890 304751 Issue 1 4 20 Rev 0 The PSE can perform translation of the addresses in Call Request packets for X 25 and X 75 ports The called and calling addresses in the packet may be independently translated as the call enters or leaves the PSE
200. priate files on the local client node an entire group of files can be copied to the remote server node in one operation Section 5 3 4 once again contains the full syntax but a brief example would be Copy a atL a123 X890 304751 Issue 1 L 1 Rev 0 This pattern matches all data files all boot files and all load files The destination is simply node 123 drive a Enhanced pattern matching is supported from Version 8 2 L 1 3 Self Extracting Compressed Load Files A full set of load files would add up to more than 2 Megabytes of data To save and rationalise disk space a compression scheme was applied to the U03 type load files These files are just normal files to the file system but when loaded they extract themselves to their original size as well as performing a 32 bit CRC check to validate the extraction Compressing the files also reduces significantly the amount of data that needs to be transferred between nodes An 8325 disk set can now be stored on a single disk drivea Driveb is thus free for dump files online backups of configuration files and downloading of new versions of software An 8425 8525 disk set is held entirely on drivea except for the Node Manager load file This is because of the number of X 25 load files supported on drivea and the need to support imported applications such as TGate L 1 4 Move Command An additional Utilities Disk Utilities screen was added to move files i e rename them This was
201. quence numbering modulo 8 will be supported The default is NO Level 2 Window Size K This should be set to match the window size required by the connected device Values are from 1 to 7 for basic sequence numbering and 1 to 127 for extended sequence numbering The default is 7 X 75 Support X 75 1980 extended sequence frame format is different from X 75 1984 and X 25 Values are NO X 75 1980 and X 75 1984 The default is NO Protocol Option If this parameter is set to 1 or 3 passive then the Level 2 code will operate in a manner consistent with X 25 CCITT If the parameter is set to 2 or 4 passive then the Level 2 code will retransmit unacknowledged frames after a link reset In passive mode 3 and 4 the PSE waits for the attached device to initiate the Level 2 start up procedure The default is 1 The PSE complies with NET2 Section 9 1 1 DCE initiated link set up and Section 9 1 4 DTE initiated DISC start link set up DCE initiated is always enabled DTE initiated DISC start is enabled with option 1 or 2 M ode of Operation This parameter determines whether the Level 2 code operates as a DCE or a DTE The default is DCE Data Link Connection ID DLCI In the case of a logical port which maps to a frame relay virtual physical port this parameter specifies within which DLCI this port s traffic will be carried over the frame relay physical interface l e it provides a mapping between the logical port number and
202. r so that you can logon when the equipment first arrives the PSE arrives programmed with a default username and password as follows username wizard password wand Once successfully logged on the terminal screen should display the Node Manager s main menu 2 2 4 Menus and Commands The PSE is configured and its operation controlled by use of menus There is a tree of menus see Apendix H starting from a Main Menu which has an option for each area of the functions provided i e Alarms Warnings Billing Configuration Routing specification Statistics Utilities Management of applications X890 304751 Issue 1 2 5 Rev 0 Selection of one of these options in most cases causes a further menu screen to be displayed and this cycle continues until you reach a final action screen For example if you select Billing then the next screen allows you to select the address of a billing collection device if you select Configuration the next menu presents a choice of configuring a node module port logical port PVC hunt group or CUG and selecting one of these options displays a menu of configuration options e g edit delete and so on All the menus used by the Node Manager follow the same format and all operate in a consistent fashion The top and bottom lines of the screen show the same information no matter which menu is being displayed The top line of the screen shows the node s number and name and the current date
203. r Handling 6 1 The Virtual DTE Facility The virtual DTE facility is a useful piece of diagnostic software used to determine network and configuration problems You can connect to it from any point in the network provided a route exists to that point from your port Every Xpress PSE comes complete with a set of virtual DTEs that should meet your diagnostic requirements 6 1 1 What It Is Used For e nter node configuration problems Virtual DTEs can be used to determine a terminal s accessibility to Xpress nodes within the network You can specify different node numbers preferably ones that are in the network to try out each route in turn If any routing problems are encountered you can use the Manager terminal on the PSE to take corrective action e Local configuration problems Problems such as parity errors local and remote echo problems and logical channel mismatches can be determined by calling the virtual DTEs on the local node e The virtual DTE mechanism is also used to access the Aynchronous Broadcast Service described in Appendix G 6 1 2 How It Is Accessed Make an X 25 call as in Section 3 5 but replace the logical port number with the one reserved for the virtual DTEs See below and replace the sub address by a number in the range 0 2 The following logical port numbers are reserved for virtual DTEs 9ss9 Where ss indicates the slot numbers on which the addressed virtual DTE resides A value of 00 addresses
204. r any User Facilities or User Data a very small amount that the DTE wishes to send X890 304751 Issue 1 A 7 Rev 0 2 The DCE uses the specified destination address to route the packet across the Packet Switching Network to the remote destination DCE 3 The remote DCE forwards the packet to the called DTE as an Incoming Call packet choosing the lowest logical channel number which is available at that link 4 The called DTE accepts the request for a virtual circuit by sending a Call Accepted packet toits DCE 5 This packet is passed back across the network and sent to the calling DTE as Call Connected The two DTEs may now exchange data across the SVC DTEs exchange data in Data and Interrupt packets over a PVC or established SVC Various other types of packet are used for flow control or to reset the circuit if a problem arises Toclear the SVC either of the DTEs may send a Clear Request packet to its DCE The packet is forwarded to the remote DTE as a Clear Indication which the DTE acknowledges with a Clear Confirmation The Clear Confirmation packet is then passed to the DTE which originated the clear down The SVC has now been removed and the logical channels may be allocated to other virtual circuits A 2 3 1 The Fast Select Facility The procedure described above indicates that an SVC must be fully established if a DTE wishes to send more than a few bytes of user data However X 25 provides the Fast Select facility w
205. r boards by examining the Alarms screen If you can t log on then the boot up could have failed for any of the following reasons Nodisks in the drives Thedisks are write protected or incorrectly formatted Thedisks are in the wrong drives Thedisks contain no load files TheUM failed its hardware diagnostics indicated by an F onthe board s display It will be necessary to contact your supplier for advice in this case X890 304751 Issue 1 2 2 Rev 0 2 2 Introduction to the Node Manager The Node Manager has four different access methods only one of which should be used at a time The presentation of the manager screen during each of the methods of access is very similar any differences are identified in the descriptions below 2 2 1 Access to the Node Manager e Fromthelocal terminal Mini PAD The Mini PAD provides a Triple X PAD type interface to the operator and to the network The standard method of access is via a VT 100 terminal plugged in to asynchronous port 1 of the UM see the relevant Installation Guide for port definitions When the PSE is running and the terminal is switched on press RETURN on the keyboard to display the Mini PAD prompt Typing LOCAL RETURN or L RETURN see Section 2 2 2 connects you to the local Node Manager You are then prompted to login If you provide a valid username password pair the logon will be successful e g Enter username wizard e Fromthelocal terminal o
206. r needed These circuits are established by the connected X 25 X 75 devices making call request call accept exchanges addressing each other by means of the logical port number of the Xpress port to which they are connected No further internal configuration is required The way in which the logical port number is incorporated in an X 121 address is described in Section 3 5 1 below e Permanent Virtual Circuits PVCs which are allocated for a period of time They are always ready for use in the same way as dedicated physical circuits but do not consume network bandwidth unless transferring traffic PVCs cannot be configured for application ports Unlike SVCs which are set up by attached X 25 X 75 devices PVCs must be explicitly configured on behalf of those devices as explained in Section 3 5 3 3 5 1 Xpress Internal Addressing In order to make a call between two attached devices the calling device must send a call request into the node specifying the logical port address of the called device in the called X 121 number field of the call request packet Internally Xpress uses 11 to 14 digit X 121 numbers of the form 12345 6 7 89 10 11 12 13 14 DDDDNNNLLL LS S Digits 1 to 4 are the Data Network dentification Code DNIC which for calls which stay within a single Xpress Network are always set tothe escape value of 1100 The use of DNI Cs to support gateways to other networks is covered in Section 3 9 Digits 5 to 7 are the Node Num
207. raises Control to the ISDN TA to get it to dial the remote TA Once this has been done and the trunk is up the call is forwarded to node 3 via the dial up trunk and re established with the host If the Auto Reroute feature is configured on the dial up trunk the user s call will be periodically internally cleared and re established in an attempt to move it back to the primary link should it have come back into service Should this happen the software will notice that the dial up trunk is no longer required and will drop Control tothe TAs thus clearing the ISDN call This actually happens after a 120 second delay in case another call should turn up requiring the link If the user clears the call the modem link will also be broken by the software after a 120 second delay once it has decided the modem link is no longer required X890 304751 Issue 1 K 2 Rev 0 K 2 Operation and Signalling K 2 1 General The dial up software uses the V 24 or V 11 circuits defined below to control the attached device It should be noted that incoming calls can only be detected if the Monitor Test Ring I ndicate option on the Physical Level configuration screen is enabled It should also be noted that dial up operation is not available on V 35 or V 36 interfaces which cannot control the required signals If the dial up operation is fully symmetrical i e both ends of a trunk are set to be dial up then an X 25 call from either end of the trunk can initiate d
208. red to specify which drive to verify e g Verify a this will start the verify operation off for drive a The section titled Automatic Disk Verification which follows provides more information on the reason for doing a verify operation 5 3 8 Automatic Disk Verification Faulty disks or disk drives can prevent the PSE from operating correctly Such faults may only become apparent after an important operation fails For example the PSE may fail to re load a crashed card if the system disk has become corrupt To try and detect such errors as soon as possible the PSE regularly checks the system disks Disk verification is performed daily on the anniversary of the system being powered up At this time the node manager verifies the disks present in each drive Verification takes approximately five minutes per disk Events are generated to indicate that verification has started and that it has finished successfully An alarm is generated if a fault is discovered on either disk X890 304751 Issue 1 5 9 Rev 0 5 4 Dump Utilities This menu lets you select the dump utility function you need A dump file is generated automatically after a module has crashed unless disabled Dump files contain useful debugging information for the Cray engineer They are stored on the disk in driveb Dump files are identified as core b u where b bay number and u module number e g core 0 9 Only one file is kept for each module If the same module were to cr
209. reload using the Configuration Module Configuration Restart Module SCreen When a module crash occurs an alarm event is raised to notify the operator The operator can then follow the progress of the board as it recovers by using the Configuration Node Configuration Node Status screen For example after a software failure the following module state changes are to be expected Unknown Software Error Dumping Loading Operational If the failed module is a XIM all calls across that module are lost If the failed module is a UM any system management calls e g remote printing or billing will be lost However they will automatically re establish once the UM is operational again Whilethe UM is not operational existing calls across any XIM are unaffected but no new calls are allowed The on line record of events is lost when the UM crashes X890 304751 Issue 1 6 4 Rev 0 6 3 Dump Files A dump file will be generated automatically if a module crashes provided that the Dump facility is enabled Only one dump file is kept for each module so it should be copied to diskette immediately in case the module crashes again and causes the dump file to be overwritten A full description of dump files and how to use them is given in Section 5 4 X890 304751 Issue 1 6 5 Rev 0 X890 304751 Issue 1 6 6 Rev 0 Appendix A X 25 Packet Switching and Frame Relay A 1 Introduction Packet Switching is the transport of blocks or packets of
210. request The Recognised Private Operating Agency RPOA facility enables a call to be routed through a specified sequence of transit networks The RPOA overrides the normal called address routing See Appendix A for more information 3 9 1 Internetworking DNIC IDNIC When interfacing to other networks the Internetworking DNIC is used as the DNIC for Xpress nodes It does not replace the Internal Xpress Escape DNIC 1100 The Internetworking DNIC is used in the following ways e When the RPOA selection matches the Internetworking DNIC a node will remove the RPOA selection from the incoming call request If there are no more RPOAs in the call request then the called address will be used for routing the call This works for incoming call requests received at Trunks as well as at Ports e Thelnternetworking DNIC is used as the Xpress Network identity i e TNICs and CNIC for calls through an X 75 gateway port The DNIC must be configured for X 75 gateway ports to work If it is NULL then all X 75 calls will be cleared The Internetworking DNIC is configured on the configuration node configuration edit node configuration SCreen X890 304751 Issue 1 3 44 Rev 0 3 9 2 Calls Toa PDN Routing to a gateway port is achieved by associating the network s DNIC the first four digits of an X 121 address or RPOA Selection in a Call Request with a list of up to three gateways to the network For example Node Number Port Number Primary gat
211. restriction bit 4 a protocol identifier is recorded for the call bit 3 set if the call was redirected bit 2 set if the call was deflected bit 1 set if the call used network data integrity bit 0 for future expansion Coding Ox0047 Ox1047 integer integer integer integer integer integer integer integer bcd bcd bcd bcd integer integer integer boolean boolean boolean boolean boolean boolean boolean boolean boolean boolean boolean boolean boolean boolean Table C 1 X 25 and X 75 Billing Information Record X890 304751 Issue 1 C 2 Rev 0 Byte Offset Size Contents Coding 50 word CUG selection see Note 2 bcd 52 word CUG OA selection bcd 54 byte recorded NUI length integer 55 byte 16 NUI integer 71 byte recorded user id length integer 72 byte 16 user id integer 88 byte 4 protocol identifier see Note 1 integer 92 word transit delay in millisecond ticks see Note 4 integer 94 byte clearing cause integer 95 byte clearing diagnostic integer 96 byte redirection deflection reason integer 97 byte unused padding byte 98 long call duration in 50 ms ticks integer 102 byte called packet size see Note 5 X 25 103 byte calling packet size see Note 5 X 25 104 byte called window size X 25 105 byte calling window size X 25 106 byte called throughput class see Note 6 X 25 107 byte calling throughput class see Note 6 X 25 108 long total called data bytes integer 112 long total calling
212. rk address of the application s manager This can be any valid network address up to 15 digits in length and need not be the address of a port on the local node The address is not restricted to the Xpress physical address format provided that the appropriate addressing tables have been suitably configured See Section 4 5 Username The username to be supplied to the application when logging in This is free format text up to 16 characters in length If no username i e an empty string is configured for this application then will be displayed in this field Password The password to be supplied to the application when logging in This is free format text up to 16 characters in length For security this field will always be displayed as Xxx thus it is not obvious if an application does not have a password configured The password may contain any characters except for RETURN X890 304751 Issue 1 2 8 Rev 0 For security the Node Manager will not echo your keyboard input when you type in the password Also the Node Manager will prompt you to enter a new password twice for verification If you do not have wizard permissions then the Node Manager will prompt you for the password if you wish to change a password or delete an entry 2 3 2 Access to the Management Screens of an Application The Application Management screen provides the Login command which you select to log into the manager of an application When you select the
213. rminal A client no longer wishing to receive broadcast data should then simply disconnect the call in the normal manner X890 304751 Issue 1 G 3 Rev 0 A client who wishes to access a server can be located anywhere in the network provided that a call can be made to the required server address then the broadcast data can be picked up The server is best located as close as possible to the majority of clients rather than close to the host For example where a maximum of 32 clients wish to access broadcast data from a host connected to an adjacent node if the server used is on the host s node then 32 copies of the host data will be sent via 32 calls across the inter nodal trunk However if the server is located on the client s node then only one copy of the data will transit the trunk as the duplication will be done locally The X 25 ports can be configured at levels 1 2 and 3 as required and as described in Section 3 4 The network will take care of any buffering and or fragmentation of data required together with facility mapping etc Any call user data or facilities settings present in any client call other than the first call which wakes up the server are simply absorbed by the server and not passed on See Section G 2 2 for details G 2 2 Server Access to the Host The X 25 call to the host is only established once the first client connects to the server The call remains in place while there are any clients connected and is
214. rmissions is given in Appendix D 4 6 3 Setting up CUGs CUGs and their members are identified by two index numbers e A global index number identifying a CUG within the network e A local index number identifying ports on one node which belong to the same CUG These two indices are mapped together As an example suppose that you want to set up a CUG covering nodes 1 and 2 called CUG 156 156 is the global index number On node 1 designate say CUG 1 to be the local CUG so map 1 to 156 On node 2 you may want CUG 89 to be the local CUG so map 89 to 156 The network wide CUG 156 now consists of CUG 1 on node 1 and CUG 89 on node 2 Note that the local indices on each node do not have to be the same but the global index number must be the same on each node if the CUG is to be regarded as spanning more than one node 4 6 4 Configuration of Local to Global Indices This is always the first task when setting up CUGs 1 Select the Configuration Closed User Group Configuration Map Local CUG Indices to Global Indices Menu 2 To create a new mapping select the Create mappings option Enter an index number range 1 99 Since this is a new mapping this index number must not be in use already X890 304751 Issue 1 4 24 Rev 0 To edit or delete a mapping select the required option then enter the index number to be edited or deleted 3 Enter the global index number range 1 65535 i e the number of the CUG to which this index be
215. rrange things so that the port which supplies the clock is the DCE Also the two ends must have the same window size and sequence number setting 4 3 1 4 Network Level Parameters These parameters control the Packet Level operation over a trunk They are a subset of the parameters configured at X 25 ports described in Section 3 4 4 e Logical Channel Number Boundaries Only Two Way logical channels are supported at trunk ports The valid range for LCNs is 0 1023 If no channels are to be allocated for a trunk port then the lower and upper channel boundaries of the Two Way group are set to NONE As with X 25 ports one UPM3 XIM or SP XIM combination supports a maximum of 512 channels shared between its four or six ports The UPM4 SP XIM can support up to 1024 channels e Extended sequence numbering If this parameter is set to YES then the PSE will support extended packet sequence numbering modulo 128 over the trunk Otherwise X890 304751 Issue 1 4 5 Rev 0 DasiCc packet Sequence numbering MOAUIO o Wiii DE Supported Ine default is NO e Window Size Set this to match the Level 3 window size required by the connected trunk port Values are 4 7 for basic sequence numbering and 4 127 for extended sequence numbering with defaults of 7 e Trunk Port Mode This field selects whether the port operates as a DCE or DTE If this trunk port is configured as a DCE which by convention supplies the line clock the connected trunk po
216. rt on the other PSE must be configured as a DTE set for External clocking and vice versa The default mode is DCE e Backward Compatibility Mode This parameter must be enabled if the trunk connects to a node running an earlier version of PSE software The parameter allows you to select the software version with which you need tointerwork By default backward compatibility is disabled e Dial up Operation and Dial up Timeout These two parameters control the behaviour of the trunk dependent on whether the link to the attached device is permanently available i e leased line or permanently connected modem link or dialled i e dial up modem or ISDN TA link See Appendix K for full details of these two types of operation e Auto Reroute Interval This parameter if set to a value greater than 0 represents the number of minutes after which all calls transiting the trunk that are not using the trunk as their optimum choice will be cleared and automatically rerouted transparently to the caller Calls are only considered to be using the trunk optimally if this trunk is the primary next hop to the destination node in the routing table Calls which were originally established on a primary next hop trunk can get routed down a secondary or tertiary if problems occur on the primary such as congestion high error rate or failure This facility allows such calls to be re instated on the primary This is particularly valuable when us
217. rts to influence routing decisions The mechanism is simpler than used in Congestion Monitoring When a manager defined limit is reached the port is closed for a configurable period The error rate is determined by the number of REJ frames on a port thus it reacts to errors on the transmit and receive paths The current valueis displayed on the menu Configuration Node Configuration Detailed Link Status Display This screen also displays errs instead of up if the port is closed because of high error rate In general actions to close a port should only be configured when there is an alternate route to the destination X890 304751 Issue 1 N 1 Rev 0 N 2 Parameters to be Configured This feature is configured on a per port basis using either the menu Configure Port Configuration X 25 X 75 Application Port Configuration Error Monitoring or the menu Configure Port Configuration Trunk Port Configuration Error monitoring The default values disable the feature e Error monitoring period The time over which the error count is averaged From zero disable to 9 minutes 59 seconds e Port reinstatement delay Up to 23 hours 59 minutes e Error tolerance limit The percentage of errors permitted before the port is closed and the system attempts to re route existing calls A default of 100 disabled any action X890 304751 Issue 1 N 2 Rev 0 Appendix O Configurable DNIC Until Version 9 the product always used DNIC 1100 for its
218. s reply to commands etc For more information refer to the Cray 5800 or 5x50 Operators Guide 3 9 5 DNIC Barring Table DBT The DNIC Barring Table is used to police call requests arriving at an X 75 gateway port from an external network DNIC Barring is driven by a user configurable table The table is set up using the Routing Specification DNIC Barring Table screen The DBT is organised as an ordered list containing the Calling DNIC Called DNIC and Status The table is searched from top to bottom for a matching entry If no match is found then the call will be allowed into the Xpress network Call request packets received from an external network attached to the X 75 port will have their called and calling address DNI Cs compared against each table entry in sequence until a match is found If a match is found then the Status field is used to determine the action to be taken If the Status is bar then the call will be cleared If the Status is allow then the call will be forwarded into the Xpress network If the PSE detects that the DNIC Barring Table is corrupt then all calls arriving on the X 75 gateway from the external network will be barred X890 304751 Issue 1 3 46 Rev 0 Valid characters in the Called and Calling DNIC fields of the DBT are Oto9 aspecific digit in the range 0 to9 n allows any digit in the range 0 to 9 Example DNIC Barring Table Calling DNIC Called DNIC Status a 1234 5678 allow b 1234 5nnn bar
219. s automatically following the recommendations of T1 618 Annex A LMI notified frame relay PVC and link failures are actioned immediately causing the end to end LAP B link to go down and X 25 calls to be cleared and if appropriate re routed e it is not necessary for the LAP B link to time out to recognise a frame relay link failure Figures A 8 and A 9 show how Xpress X 25 and trunk protocol packets are encapsulated within frame relay frames X890 304751 Issue 1 A 25 Rev 0 Host 1 Frame Relay Network DLCI 102 Slot 3 Port 0 Node 1 Host 2 Xpress X 25 X 75 ports multiplexed over frame relay The logical X 25 connection to Host 1 is realised by DLCI 101 over the frame relay link on slot 3 port 0 The node and host exchange X 25 packets encapsulated within frame relay frames Above the encapsulation software both the node and host think they are directly connected via X 25 i e they are acting as FR PADs X 25 L3 packet IV toO OHO LAP B frame A LAP B frame and X 25 level 3 packet are encapsulated in the frame relay frame Figure A 8 X 25 Encapsulation X890 304751 Issue 1 A 26 Rev 0 Node 2 Node 3 Frame Relay Network Slot 3 Port 0 Node 1 Node 4 Xpress trunks multiplexed over frame relay The logical trunk from node 1 to node 2 T0002 is physically realised as DLCI 102 on the frame relay port on slot 3 port 0 A similar mapping exists for the other trunks and nodes but is not shown for Cl
220. s containing only trunk ports are intended for inclusion in the routing table as inter node routes In large configurations where there may be a number of trunks to the same destination the trunks may be collected into a trunk group The trunk group can then be specified as one of the entries in the routing table Any calls subsequently forwarded over this route will use the trunk selected by the trunk group mechanism thus spreading the load over all the individual trunks Trunk group addresses are in the range T8000 T 8999 Members of trunk groups must themselves be trunk ports Congestion Monitoring and Control can often be beneficial when configured on ports belonging to trunk groups in order to balance the utilisation more evenly between the trunks See Appendix M for more details X890 304751 Issue 1 3 43 Rev 0 3 9 X 25 X 75 Gateways Ports in an Xpress PSE network may be used as X 25 or X 75 gateways to other Public or Private X 25 networks for example to British Telecom s Packet Switched Service PSS Calls destined for such a non Xpress network will specify addresses that are meaningful to that network rather than to the Xpress PSEs The Xpress PSEs will route such a call toa Public Private Data Network PDN gateway if e The DNIC of the called address is not the Xpress Escape DNIC 1100 The first four digits of the called address are taken as the DNIC of the destination network e An RPOA selection is present in the call
221. s nothing there on which to record them This may also result in inconsistent information being recorded on the disk 3 7 4 X 25 Data Link Down Assuming the XIM is running correctly the most likely cause of this isa physical problem with the cable or the level 1 configuration 1 Ensure that the clocking set up is consistent with the requirements of the PAD i e that one end is clocking not both or neither 2 Fromthemain menu gotothe Statistics Display Port Statistics physical level screen Type in the logical port number of the bad port and check that the status of the control lines matches those expected for the connected device Serious line errors will be shown up by high counts in the other fields 3 If all is well press PF3 and enter the Statistics Display Port Statistics frame level screen which will show you what the two ends of the data link are doing 3 7 5 X 25 Call Failed There are two ways the call can fail assuming that the level 2 data link is up 1 Thecall is cleared immediately 2 The call disappears and is cleared after a delay There are many possible reasons for 1 and a detailed description of all the things that can go wrong is out of the scope of this document If none of the reasons suggested below applies contact your supplier X890 304751 Issue 1 3 39 Rev 0 e Bad called address Ensure that the address used is correctly formatted It must bein the Address Analysis Table See Section 4 5
222. section briefly describes the functions of the parameters and the values to which they may be set For more detailed information about X 25 port configuration and particular parameter settings see Appendix A Only Network Level Parameters and User Facilities can be configured for Application Links The Frame Relay Core Level Parameters are only applicable to the physical port connected to a frame relay network interface These parameters may seem somewhat daunting to inexperienced users but most of them can be left with default settings Note that there are very few differences in the configuration requirements for X 25 and X 75 ports Differences are mentioned where they exist 3 4 1 Physical Level Parameters These govern the characteristics of the physical connection between the PSE port and the directly connected device Note that when configuring the physical level of a frame relay virtual physical port via that port s logical port number it is the physical port over which the virtual physical ports are multiplexed that is being configured E g if physical ports 0 2 0 to 0 2 5 are multiplexed over physical port 0 2 0 and hence become virtual physical ports and logical port 0024 which maps to virtual physical port 0 2 4 is edited at the physical layer then it is physical port 0 2 0 which is changed e Clock Source If this parameter is set to External the PSE will use the line clock X890 304751 Issue 1 3 10 Rev 0 suppiled
223. self by means of the entry of a User ID and password the ACS provides a choice of available services After selecting a service the user s call is then transferred from the ACS to that service Should the connection to that service fail the user may be re connected to the ACS to select an alternative without having to repeat the logon procedure Full details of this procedure are given in the ACS User Guide An example network is shown in Figure E 1 X890 304751 Issue 1 E 1 Rev 0 ACS SLAVE 1 COMMUNITY A ASYNCH USERS COMMUNITY A DCX DCX USERS NETWORK PAD X GATE WizzNet DNIC 3001 0002 0003 j XPRESS NODE 2 0001 UNPROTECTED COMMUNITY B ASYNCH USERS ASYNCH USERS 0001 EA XPRESS XPRESS 0o04 a NODE 1 NODE 3 0002 oe 0003 PSS 3 0003 0004 wae 2342 Te 0002 MASTER HOST 1 HOST 2 HOST 3 P p EE f e 2 Figure E 1 Example ACS Network X890 304751 Issue 1 E 2 Rev 0 E 2 Support for ACS In order to support the functionality described above Xpress uses a number of standard and non standard CCITT X 25 User Facilities together with the standard Xpress address translation facilities as follows Incoming Calling Address Translation This mechanism is used on user ports which are to be subject to the control of the ACS For all such ports a translation is set up to replace some or all possible calling addresses with the address of the ACS Thus any call the user makes to a protected service ma
224. sing RETURN has no effect check the terminal connection tothe UM Thescreen should display the message X 25 PAD port 0 followed by the Mini PAD s prompt Typing help RETURN or h RETURN will give a list of the Mini PAD commands CON lt X 121 address gt lt CR gt HELP lt CR gt LOCAL lt CR gt REMOTE lt node number gt lt CR gt SET lt par gt lt par value gt lt CR gt SET lt par gt lt par value gt lt CR gt PAR lt CR gt The X 3 parameters supported are 2 3 4 9 10 13 14 and 15 The commands HELP LOCAL and REMOTE can be specified by their first letters H L and R e CON attempts to establish an X 25 call e LOCAL connects to the local Node Manager e g L RETURN e REMOTE connects to a remote N ode Manager e g R nnn RETURN connects to node nnn e SET sets the values of one or more X 3 parameters SET 3 2 2 0 X890 304751 Issue 1 2 4 Rev 0 e SET is similar to SET but the Mini PAD also confirms the values that have been set SET 3 2 3 0 PAR 3 2 3 0 In this case parameter 3 is first set to 2 and then to 0 SET and SET can only change the parameters that are supported If the user tries to change a different parameter the Mini PAD replies with INV SET 3 2 16 1 INV PAR 3 2 e PAR lists all parameters from 1 to 22 2 2 3 Logging on to the Node Manager Access to the Node Manager is password protected Once logged on you can create your own usernames and passwords Howeve
225. state the Node Management loading system needs help from the operator Examples that cause the module to enter this state are cannot find or access module load image on disk module reload flag not set Please wait The management system is performing some transient operation on the module e g while restarting it When a module crash occurs or a new module is inserted into the node etc the management system automatically takes any action necessary After a software failure the following module state changes are to be expected Operational Unknown Software Error Dumping Loading Operational When a new module is inserted the following state changes are to be expected Unknown Idle Loading Operational 5 11 2 Detailed Link Display This screen displays detailed information for every logical port in the node The first line of the display provides totals of the number of links up PVCs and SVCs connected for the node For each logical port the following details are displayed Description the port s name Port State configured state online offline Or out of service X25 Layer 2 state up down or errs PVC number of PVC calls connected SVC number of SVC calls connected UTLs utilisation level ERRs error rate If the module corresponding to a link has failed then the Layer 2 protocol state is displayed as Error rateis shown as 0 unless the trunk link is on offline and up with Error Monitoring enabled
226. such as the Cray 8160 Group LowLCN High LCN PVC 512 527 Incoming 1000 1008 Two Way 1009 1487 Outgoing 3072 3079 Here all 512 available LCNs have been used up by allocating 16 PVCs 9 Incoming Only circuits 479 Two Way circuits and 8 Outgoing Only circuits No other ports can be configured on the XIM because there are no LCNs available for them e Default Maximum Packet Size This is the maximum data packet size to be used for all calls through this port unless a different value is negotiated during call setup Values are 16 32 64 128 256 512 1024 2048 and 4096 with a default of 128 Note use of large packet sizes on a card with only 1 Mbyte of memory may result in a shortage of packet buffers e Maximum negotiable packet size This is the maximum packet size which can be negotiated during call set up using flow control parameter negotiation Values can range from 16 to 4096 octets The default is 256 octets e Mode of operation This parameter determines whether the PSE port operates as a DCE or a DTE at the packet level The default is DCE X890 304751 Issue 1 3 17 Rev 0 e Profile identifier This parameter determines which protocol the PSE supports at the packet level over the link The protocols are CCITT X 25 1980 PSS CCITT X 25 1984 Tymnet Telenet CCITT X 25 1988 Pl CCITT X 75 1980 P2 CCITT X 75 1984 P3 The PSE also provides three spare profiles These are included so that
227. ted by the trunk group mechanism thus balancing calls over all of the individual trunks 4 4 6 Hop Counts Each entry in the routing table also includes a maximum projected hop count field The hop count is used to prevent a call from circulating endlessly within the network almost certainly due to a wrongly configured routing table The maximum projected hop count specifies the maximum number of nodes a call is allowed to traverse en route to that destination Once a call has passed through this number of nodes it will be automatically cleared X890 304751 Issue 1 4 17 Rev 0 4 5 Addressing 4 5 1 Address Analysis Address Analysis allows any format addressing scheme to be imposed on an Xpress network provided the address is configured in the Address Analysis Table Each node has its own Address Analysis Table The Address Analysis Table is used to analyse the called address and to translate the logical address into an Xpress Internal Address for routing The called address in the Call Request Packet remains unchanged Address Analysis is driven by a user configurable table The table is set up using the routing specification address analysis table screen The Address Analysis Table is organised as an ordered list of Match Address and Internal Address pairs All Call Request Packets received at the node have their called address compared against each table entry in order from the top until a match is found If a match is found th
228. ted with an error message but without giving you the option of overwriting any existing application file In this case you must first either delete the existing application or remove the existing individual files from the system disk and then select the Install command for a second time So for example if you wish to install a new version of an existing application then you must first delete the existing application X890 304751 Issue 1 5 11 Rev 0 5 5 3 Deletion The Delete command allows you to remove a previously installed application from the PSE Thescreen does not prevent you from deleting applications which are in use When you select the Delete command you will be prompted to give the entry number of the application An angle bracket gt will be displayed against the chosen application indicating that it has been selected You can now choose to delete the application by selecting the PF1 key or abort the delete operation by selecting the PF3 or PF4 key If PF1 is entered the application is deleted and the list of installed applications is updated on the screen The Node Manager removes from the system disk the files used by the application The database files applic data and novid data on the system disk will be updated to remove details of the deleted application Note that the Node Manager will not remove any database information or load files which are shared with other installed applications 5 5 4 Expand T
229. terface to the Xpress software on the UPM via Library functions which Cray provides X890 304751 Issue 1 J 3 Rev 0 Communications Interface s if present ACM Hardware no processors AAAA AAAA AAA AAAA PD POET ELS UPM Memory Map APPLICATION ee ee ee ee eee ee a a a a a ee Fg an E ae a Dye 71 m i te are ee e A a A l Restof OPERATING the Xpress SYSTEM Kernel Software i eee ENE he an et te pel Xpress memory partition lower 2Mbytes Xpress inter card B Bus Figure 1 UPM Co Resident Application X890 304751 Issue 1 J 4 Rev 0 Intelligent ACM UPM XPRESS KERNEL SOFTWARE if present Figure J 2 ACM Application on a Card with One ACM Processor Intelligent ACM UPM Comms Lines APPLICATION B a 3 B S if s XPRESS presen KERNEL Processor No 1 SOFTWARE EEA AAAA AAAA Comms Lines APPLICATION i 3 l present oe Processor No 0 Figure 3 ACM Application on a Card with Two ACM Processors X890 304751 Issue 1 J 5 Rev 0 J 5 Application Programming Interface API J 55 1 Overview Applications access the Xpress communications and management services by making function calls to the Cray supplied Library which in turn invokes functions provided by the Xpress Kernel software These function calls emulate a subset of UNIX system calls e g connect listen send recv and a subset o
230. terion used to represent the degree of congestion present A new value is computed every 15 seconds and displayed as a percentage on the menu Configuration Node Configuration Detailed Link Status Display Transient spikes or dips in the value are smoothed by averaging each new value with its predecessor The utilisation is continuously measured irrespective of whether any of the congestion control actions described below has been set up to take effect on the port e Comparing utilisation against configured thresholds On a per port basis for each of the four priority classes two utilisation threshold levels can be configured one for call refusal and one for call clearing Every 15 seconds the system compares the current link utilisation against each of these configured thresholds and takes action if appropriate e Call Refusal Setting Call Refusal thresholds is a straightforward way of limiting access to a congested trunk or link While the measured utilisation level exceeds the Refuse threshold configured for a particular priority class further calls of that priority are refused access to the trunk or link The system will attempt to establish the refused calls via an alternative route if one exists such as the secondary or tertiary next hop trunk port e Call Clearing Bumping Refusal thresholds can be supplemented with the more severe measure of internally clearing existing calls This comes into effect only if there is at
231. the test pattern generator See Appendix F for more details of the V 54 test loops If the module supporting this port is not operational statistics will not be available 5 10 3 Frame Relay Core Level Statistics This screen displays the frame relay core level statistics for a specific frame relay physical port The core level statistics are those concerned with the basic data transfer role of the physical frame relay interface and include information on the activity and congestion status of the link as well as counts of frames and bytes transmitted and received The time since the statistics were last reset is also displayed If the module supporting this port is not operational statistics will not be available Note that if this screen is selected for a logical port that maps to a frame relay virtual physical port the statistics displayed will be for the physical port over which the virtual physical port is being multiplexed 5 10 4 Frame Relay LMI Statistics This screen displays the frame relay Local Management Interface statistics for a specific frame relay physical port the LMI runs over DLCI 0 on the physical frame relay interface and the statistics include information on the protocol transactions exchanged over this DLCI by which means the node ascertains the configuration and overall reliability of the link The time since the statistics were last reset is also displayed If the module supporting this port is not operatio
232. their severity is fixed X890 304751 Issue 1 5 15 Rev 0 5 7 1 Alarms and Warnings When an event occurs it is logged to the printer If the event is an alarm or warning and events are being handled locally its details are also maintained on line to allow the operator to examine them Alarms and warnings are displayed on separate screens to save the operator having to sort them out The same display format and commands are used on both screens The counts line at the bottom of the VDU screen is updated to show that a new alarm or warning has occurred and this is accompanied by an audible bleep to notify the operator When there are no alarms or warnings this line displays OK The system can hold a total of 100 alarms and warnings If more come in the oldest event is overwritten by the new one but the operator is warned of this before it occurs When a new alarm occurs you should check it on the Alarms screen Ten events termed a page can be displayed on the screen at atime Once on the screen if there is more than one page full you can use commands to move backwards or forwards or move back to the first page The newest events are always displayed first If you are on the appropriate event screen and the new count for that screen shown on the bottom line of the VDU increases you can move to the first page and find out what the new event is The page of alarms or warnings shows the event status when the event was raised whi
233. then a single entry of nnn will cause all X 25 calls not destined for this node to be routed down that one route The Routing Table is set up in the same way as the Call Address Translation tables i e using the Routing Specification Routing Table Create Entry Screen 4 4 4 Routing the Call If the call is destined for this node then it is simply routed to a local X 25 port If the call is for a remote node then the routing table is checked to find the trunk that represents the primary route toward the destination The call is forwarded along that trunk and a response awaited This process is repeated at each node en route to the destination and the call is connected when a Call Accept packet returns over the virtual circuit Ifa X890 304751 Issue 1 4 16 Rev 0 Call IS not accepted along tne primary route trom a noge then It iS re triea along the secondary route and then if necessary the tertiary route until it is successfully connected Thus on a busy or damaged network it is possible for a call attempt to try alternate routes and backtrack a number of times before the connection is successfully established 4 4 5 Using Trunk Groups in the Routing Table In large configurations with a number of trunks to the same destination the trunks may be collected into a trunk group See Section 3 8 which can then be specified as one of the entries in the routing table Calls subsequently forwarded over this route will use the trunk selec
234. throughput class which is conveyed to the remote destination Xpress does not guarantee any class of throughput Xpress does not constrain the throughput classes to be the same for each direction of data transmission Flow control parameter negotiation this is the negotiation of window and packet sizes The minimum and maximum packet sizes negotiable are 16 and 4096 octets respectively The minimum and maximum window sizes negotiable are 1 and 7 respectively or between 1 and 127 if extended sequence numbers are selected Flow control parameter negotiation is carried out locally and Xpress allows different values of window packet sizes at the two ends of a call This means that Xpress provides packet fragmentation re assembl y combination as necessary The caller s requested parameters may be indicated to the called party so that it is possible for both the called and calling parties to negotiate the same window and packet sizes Throughput class negotiation this negotiation takes place between the calling and called parties Xpress does not guarantee any class of throughput Incoming Outgoing calls barred these facilities are implemented by means of the one way logical channel outgoing incoming facilities X890 304751 Issue 1 A 12 Rev 0 11 12 13 14 xw o 15 16 17 Closed User Group Xpress supports the basic format of CUG selections An Xpress network can support 65535 CUGs An Xpress
235. tic Codes B 1 B 1 Clearing Causes B 1 B 2 Resetting Causes B 2 B 3 Restarting Causes B 3 B 4 X 25 X 75 Diagnostic Codes B 4 C Billing Information C 1 D Closed User Group Call Permissions D 1 D 1 Example Network D 1 D 2 Call Permissions and Prohibitions D 3 D 2 1 CUG 1 Permissions D 3 D 2 2 CUG 2 Permissions D 4 D 2 3 Permissions for a DTE which is nota CUG Member D 5 X890 304751 Issue 1 0 9 Rev 0 mmm FF WN EF E 4 nnmnnn n BWNEFE ACS Support The ACS Support for ACS mplementation E 3 1 Call Deflection amp Call Deflection in Data Transfer E 3 2 Call Deflection Referral E 3 3 Pseudo Facility Format Xpress Port Configuration E 4 1 ACS Port Configuration E 4 2 User Port Configuration E 4 3 Host Ports E 4 4 NMC Port V 54 Modem Test Facilities ntroduction Modem Test Loops Test Pattern Generator Signals and Cables The Broadcast System ntroduction Using a Single ABS Server G 2 1 Client Access to the Server G 2 2 Server Access to the Host Using Multiple ABS Servers Providing Morethan one Broadcast Service Capacity and Performance G 5 1 Sizing G 5 2 Buffering G 5 3 Throughput Diagnostics and Error Handling Glossary MMI Tree ntroduction Xpress PSE Applications Overview J 1 1 Native Applications X890 304751 Issue 1 0 10 E I E 1 E 4 E 4 E 5 E 6 E 6 E 7 E 9 F 1 F 1 F 2 F 3 F 4 G 1 G 1 G 3 G 3 G 4 G 8 G 14 G 16 G 16 G 16 G 17 G 18 H 1 Rev 0 j t 2
236. ting the level 2 SABM UA and level 3 Restart Confirm exchange on the link However the software does act on Indicate going low treating this as a link failure and will drop Control in response X890 304751 Issue 1 K 5 Rev 0 X890 304751 Issue 1 K 6 Rev 0 Appendix L Remote Software Download L 1 Version 8 Features Version 8 of the Xpress PSE software introduces a number of features which together facilitate remote software download Each feature is briefly described in the following sections L 1 1 Remote File Operations The Node Manager software now includes a file server facility which allows a client i e another node to access the disk system of a remote node Files can be copied to or from the local client node to the remote server node Files are copied using the Utilities Disk Utilities File Copy screen This command has been extended to support remote file operations via a change in the syntax of a filename Section 5 3 4 entitled File Copy describes the command in detail Briefly file s can be copied remotely by including the node number in the filename For instance the filename a123 help data specifies the help text database file on node number 123 Single file remote file operations are supported from Version 8 1 L 1 2 Enhanced Pattern Matching To make it easy to copy a number of files from a local node to a remote node the file copy pattern matching was enhanced from Version 8 2 By specifying the appro
237. tion which concerns data transmission at 48 kbps using 60 108 kHz group band circuits CCITT Recommendation which concerns modems for synchronous data transmission using 60 108 kHz group band circuits CCITT Recommendation which concerns loop test devices for modems Virtual Circuit CCITT Recommendation which concerns User Facilities CCITT Recommendation which concerns PAD facility in a PDN Rev 0 X 21 X 21bis X 25 X 27 X 28 X 29 X 5 X 121 XIM XPAM Xpress XRMC XSAC CCITT Recommendation which concerns the general purpose interface between DTE and DCE for synchronous operation on PDNs CCITT Recommendation which concerns its use on PDNs of DTE which is designed for interfacing to synchronous V series modems CCITT Recommendation which concerns the interface between DTEs and DCEs which operate in packet mode CCITT Recommendation which concerns the electrical characteristics for balanced double current interchange circuits for general use with integrated circuit equipment identical to V 11 CCITT Recommendation which concerns the DTE DCE interface for a start stop mode DTE accessing the PAD facility on a PDN CCITT Recommendation which concerns the procedures for exchange of control information and user data between a packet mode DTE and a PAD CCITT Recommendation which concerns packet switched signalling system between public networks providing data transmission services CCITT Recommenda
238. tion which concerns the international numbering plan for PDNs X 25 Interface Module a type of ACM having 4 serial ports X 25 Physical Access Module A family of Packet Switching Exchanges using the same operational software the software itself 8325 equivalent of UPM UM card combination 8325 equivalent of SP XIM SAM also called SAC X890 304751 Issue 1 H 5 Rev 0 X890 304751 Issue 1 H 6 Rev 0 Appendix MMI Tree 1 1 Introduction This Appendix lists the Xpress MMI Tree The characters in square brackets are used to select the desired menu Section no MAIN MENU 2 2 3 A larms control 5 7 1 W arnings controls 5 7 1 B illing 5 8 C onfiguration N ode configuration N ode status display 5 11 1 E dit node identity 3 9 1 4 2 amp 5 6 C hange state of all ports on node S Jummary link status display 5 11 4 Detailed link status display 5 11 2 L ink circuit display 5 11 3 MJodule configuration 3 3 E dit module parameters 3 D isplay version number 333 C hange module link states R estart module 3 3 4 amp 6 2 PO rt configuration X 25 X 75 port configuration 3 P hysical level F rame relay core level D ata link level NJetwork level U ser facilities Co ngestion monitoring E rror monitoring C hange state of port G9 G9 G9 Go Go Go GoGo e Ro ONDOARWPY w oO Corocowwi A RRRA RAD p T runk port configuration P hysical level F rame relay core level D ata link level N et
239. ts a through g that might occur is described below and illustrated in Figure M 3 and Table M 1 a After half a minute utilisation has already passed 65 so when a second host printer call arrives it is refused access and gets diverted to the secondary trunk Node 3 T32 The utilisation caused by the existing calls continues to rise stabilising at 80 b At this point a workstation call establishes and utilisation rises briefly to 89 c The Clear settings for Priorities 1 and 2 take effect because calls of both these priorities are present and start to clear lower priority calls to reduce the congestion The first call cleared is the printer job This gets re established on the secondary trunk d Utilisation has reduced to 81 but this still exceeds the Priority 1 threshold and a further call has to be removed this time one of the terminal sessions Utilisation has stabilised at 70 after a further 2 minutes e At minute 6 the auto reroute timer on T32 expires for the second time which internally clears all 3 of the previously displaced calls so that they may re establish on the primary trunk if conditions will permit this The previously displaced terminal call is re established back on T31 but the current utilisation level of 70 is too high for the printer jobs to be re admitted so these 2 calls remain on the Secondary trunk X890 304751 Issue 1 M 12 Rev 0 Computed Utilisation 00 00 03 00 06 0
240. tual DTE ports X890 304751 Issue 1 G 7 Rev 0 G 3 Using Multiple ABS Servers As stated above the maximum number of clients that a single server can handle is 32 This is due to limitations on internal buffer space on each card However it is possible to utilise more than one server in order to broadcast data to more than 32 clients For example the server in slot 1 of a node could be broadcasting to 32 clients and picking its data up directly from a host However if the host call made by server 1 was instead directed to the address of the server in another slot for example slot 2 and the resulting host call made by server 2 directed to the host this would then allow server 2 to forward data to server 1 and also to 31 other clients of its own Therefore server 1is now a client of server 2 and the system is capable of broadcasting to 63 clients This mechanism can be extended in two ways linearly or hierarchically In the linear case the host call of server 2 would be routed to server 3 3 to 4 and soon This would result in a daisychain of servers limited only by the number of slots in the entire network In the hierarchical case each time a new server is added its host call is routed to the client address of server 2 thus using up another of server 2 s client calls but adding a new server capable of supporting 32 new clients Either of these methods can be used However the hierarchical method is likely to be simpler to confi
241. ue of the digit in the same position in the match address e g an n which is the fourth character in the substitute address takes the value of the fourth character in the match address For example Match Address Substitute address 2342 567 00123 01 1100 001 0001 001 2342 567 00123 02 1100 001 0002 002 2342 567 00123 nn 1100 001 0003 Onn nnnn nnn nnnnn nn NULL In the above table incoming addresses with sub address 01 can be forwarded to the port on node 1 with logical port number 0001 Calls with X890 304751 Issue 1 4 21 Rev 0 sub aagaress Uz are Tforwaraed to port UUUZ and any otner sup aqaresses received are forwarded to port 0003 In each case the sub address is preserved in the substitute address The final entry in the table the last match attempted traps any illegal addresses received at the port Addresses that fell unsuccessfully through all the preceding matches in the table are mapped to an invalid address which causes the PSE to clear the call immediately when routing is attempted Versions 3 1 of the software onwards assign a default incoming calling address translation such that a call request with no calling address will have the address of the calling port inserted 4 5 4 Outgoing Called Calling Address Translation OCAT Outgoing address translation is driven by two more tables per X 25 X 75 port The tables are configured and used in exactly the same way as the ICAT tables but outgoing address translation
242. urther details are given in Section A 3 3 A 3 2 How the Frame Relay Protocol Works Figures A 5 and A 6 show how the X 25 and frame relay protocols are handled within end systems and switching nodes in a network The X 25 network switches packets at Level 3 and runs the whole of the Level 2 HDLC error correcting protocol across every inter node hop The frame relay network relays frames at the core of Level 2 and runs the error correcting or procedural part of HDLC end to end between the end X890 304751 Issue 1 A 22 Rev 0 systems I netrame relay network IS generally unaware of tne Protocol s being used above the procedural part of Level 2 Network level switching Network level switching Network Network Network Network Data Link Data Link Data Link Data Link Data Link Data Link Physical Physical Physical Physical Physical Physical End system Switching node Switching node End system Figure A 5 X 25 Switching Procedural Procedural Data Link Frame Relaying Frame Relaying Data Link Core Data Core Data Core Data Core Data Core Data Core Data Link Link Link Link Link Link Physical Physical Physical Physical Physical End system Switching node Switching node End system Figure A 6 Frame Relaying The core data link layer shown in Figure A 6 is responsible for the basic Level 2 functions such as byte alignment error detection via CRC but not error correction bad frames are simply discarded transparency and f
243. various traffic types it must be arranged that each of the X 25 devices using the network generates call requests that include the appropriate Throughput Class Request value This may mean some minor re configuration of each PAD DTE or host that initiates calls into the network Alternatively the Default Throughput Class parameter can be used See below The mapping of the requested Throughput Class to a Priority Class is unaffected by the Throughput Class Negotiation process which sometimes occurs during call establishment e Default Throughput Class It may be the case that all calls entering via a given port need to have the same priority In this circumstance it is convenient to configure the Default Throughput Class parameter for that port Configuration Port Configuration X 25 X 75 Port Configuration User Facilities The configured class will apply automatically to every call request that arrives at the entry port which does not explicitly request a class Note that a PVC s priority can be set only by this means The default for this parameter is 9600 bps See Section 3 4 5 for more details e Network Data Integrity E nabled This parameter is described in Section 3 4 5 and is found on page 2 of menu Configuration Port Configuration X 25 X 75 Port Configuration User Facilities Calls are likely to be subject to a greater number of internal call re establishments when Congestion Monitoring and X890 304751 Issue 1 M 7 Rev 0
244. vel profile of the port on which they are configured In the case of a CCITT 1988 port the standard mechanism is used in both cases with a Call Deflection facility field in the facilities section of a Clear Request Packet For a 1984 or 1980 port this mechanism cannot be used as the connected device cannot supply the required facility Therefore the following mechanisms are used For Call Deflection a Pseudo facility is carried in the user data field of the Clear Request returned in direct response to a Call Request The format of this pseudo facility is given below For Call Deflection in Data Transfer the above mechanism is alright for a 1984 port but contravenes 1980 X 25 so for a 1980 port the same pseudo facility is carried in the data field of an X 29 Q bit data packet holding an X 29 1984 PAD Reselection Message E 3 2 Call Deflection Referral This uses the User Token field of the pseudo facility by copying it into the user data field of the Call Request sent to the ACS by the PSE when re establishing the user connection The ACS then extracts the token information X890 304751 Issue 1 E 4 Rev 0 E 3 3 Pseudo Facility Format The pseudo facility format is shown in the example in Figure E 2 8 7 6 5 4 3 2 1 1 h El Facility length 1 0 1 0 0 0 1 Parameter field length Call Deflection 1 1 0 0 0 0 0 0 X121 address length Selecion Facility Host service X121 address up to 15 digits m 1 1 1 0 0 0 1 1 0
245. very similar to that for X 25 ports see Section 3 4 The port configuration screens for the Physical Data link Network levels Congestion and Error Monitoring are accessed via the Configuration Port Configuration Trunk Port Configuration Menu 1 Assign Logical Trunk ports to physical ports This is done in the same way as all the other logical port assignments i e by allocating but you must prefix the letter T to Trunk port numbers e g T100 Two physical ports may be given the same LPN if one of the ports is a trunk eg LPNs 100 and T100 refer to an X 25 and a trunk port respectively 2 Follow the steps in Section 3 4 as for X 25 port configuration X890 304751 Issue 1 4 4 Rev 0 4 3 1 1 Physical Level Parameters These parameters are identical to the X 25 port Physical Level Parameters described in Section 3 4 1 What is required is for the ports at the two ends of the trunk to have compatible configurations i e one end must supply the line clock and the other end use that clock 4 3 1 2 Frame Relay Core Level Parameters These parameters are identical to the X 25 port frame relay core level parameters as described in Section 3 4 2 4 3 1 3 Data Link Level Parameters These parameters are identical to the X 25 port Data Link Level Parameters described in Section 3 4 3 Again the two ends of the trunk must have compatible configurations i e one end must be a DTE and the other a DCE The usual convention is to a
246. vice versa Before any of the other disk utilities can be used a disk must be formatted using the Format Disk option From Version 5 onwards the PSE uses high density disks and drives However it can still read and write disks produced with earlier software versions 5 3 1 Format Disk This screen lets you format a disk Disks must be formatted using this screen before they can be used Check that the disk to be formatted has any write protection removed Any data on the disk will be erased 5 3 2 Copy Disk This screen lets you perform a disk to disk copy The entire contents of the source disk are copied to the target disk Check that the disk to be copied to has any write protection removed and has been formatted correctly All data on the target disk will be overwritten A useful precaution is to write protect the source disk 5 3 3 List File Directory This screen lets you list the contents of a disk For each file present on the disk the following details are displayed file size total size of file in bytes create time time and date this file was created update time time and date this file was last updated Note If the file size value is followed by a lower case c this indicates that the file uses a compacted storage format Some of the larger files employ this format to economise on disk space The file system automatically adapts to the different formats The initial line of the display shows the number of free bytes re
247. work level Co ngestion monitoring E rror monitoring C hange state of port 3 4 8 h h h d b h ee AARABRAR OuRW X890 304751 Issue 1 l 1 Rev 0 L ogical po qh L ocal printer configuration R emote printer address specification L ogical printer configuration R emote printer address specification rt allocation Crjeate a new logical port E dit an existing logical port D elete an existing logical port L ogical physical port display P hysical logical port display Ch ange state of a logical port PV C configuration ll C reate a PVC E dit a PVC Djelete a PVC L ist all PVCs through a port H unt group confirmation ll pi C reate hunt group E dit hunt group D elete hunt group R enumber hunt group L ist hunt groups on this node C losed user group specification M ap local CUG indices to global indices S pecify CUG subscription for logical port C hange access permissions within CUG for logical port L ist CUG access permissions for logical port R outing specification R outing table configuration i I jncoming i X890 304751 Issue 1 C reate routing table entry E dit routing table entry D elete routing table entry L ist routing table entries address translation S ource calling address translation D estination called address translation Section no 3 1 amp 3 2 3 5 1 3 8 4 6 4 6 4 4 6 5 4
248. xceeded No path is available out of node Port inter node trunk is not on line Invalid CUG selection Call destination is configured to be an inter node trunk Intra node request for call re establishment invalid Intra node request for call re establishment is premature Inter node request for PVC setup is invalid End to end delivery confirmation failure Internal application is temporarily over committed Interrupt packet too long Facility field too long Intra node indication that a call is being re directed Call is being deflected Call looping detected Table B 4 continued Diagnostic Codes X890 304751 Issue 1 B 7 Rev 0 X890 304751 Issue 1 B 8 Rev 0 Appendix C Billing Information This appendix details the format of the Billing Information records generated by an Xpress PSE These records are generated when an SVC is cleared down or a PVC is taken out of service The PSE generates two records of billing information for each virtual channel one from each X 25 port An SVC s billing records can be matched together by means of the calling and called X 121 numbers and the SVC count field A PVC s records can be matched by means of the calling and called X 121 numbers and the local and remote LCls X 25 Billing Information records are 142 bytes long and the bytes may be Binary Boolean or BCD format X 75 Billing Information records are 190 bytes long Tables C 1 and C 2 show the layout of the Billing Information records The
249. y be re routed to the ACS Call Deflection in Data Transfer This is a non standard extension of the standard 1988 Call deflection facility which is configured on the ACS port and used as follows once the re routed user call has been accepted by the ACS and the user has logged on and selected the required service the ACS will use this mechanism to tell the PSE to deflect the call to the service address Further details of the implementation of this mechanism are given in Section E 3 Call Deflection Referral This is a non standard X 25 user facility which allows a failed call deflection to be referred back to the ACS for another try without forcing the user to log on again It must be configured on all user ports which use the ACS It works by accepting a User Token from the ACS during the deflection and giving the token back to the ACS if a retry is necessary The ACS takes the presence of the token to mean it can safely bypass the logon procedure Call Deflection This may be used by future versions of the ACS to deflect the original user call from the Master ACS toa slave ACS if e g the Master is busy Call Redirection This can be used to force the PSE to redirect user calls from the Master ACS toa slave if the Master is off line for some reason X890 304751 Issue 1 E 3 Rev 0 E 3 Implementation E 3 1 Call Deflection amp Call Deflection in Data Transfer These facilities are supported differently depending on the network le
250. ze of the User Facilities field to 63 bytes whereas the 1984 version allows up to 109 bytes The 1984 version introduces the CLAMN and CNIC Network Utilities The 1984 version defines timeout values and the actions to be taken when timeouts expire A 2 7 4 Calls Between X 75 1980 and X 75 1984 Ports Xpress allows calls between X 75 1980 and X 75 1984 ports To avoid violation of X 75 1980 during such a call Xpress will Clear the call if the size of a packet s User Facilities field exceeds the maximum allowed by X 75 1980 X890 304751 Issue 1 A 20 Rev 0 Remove any X 75 1984 specific Network Utilities from a packet Reset the call if an Interrupt packet holds a user data field larger than that allowed by X 75 1980 A 2 7 5 Calls Between X 25 and X 75 Ports Xpress allows calls between X 25 and X 75 ports regardless of the versions of the protocols e g between X 25 1980 and X 75 1984 ports X 75 1984 and X 75 1980 ports X 75 1980 and X 25 1984 88 ports To avoid violation of X 25 1980 or X 75 1980 during such a call Xpress will as necessary clear reset calls or remove fields from packets See A 8 and A 9 3 X890 304751 Issue 1 A 21 Rev 0 A 3 Frame Relay A 3 1 Introduction Frame relay is broadly similar to X 25 but is intended for use over the emerging fast reliable digital circuits rather than the slow unreliable analogue lines for which X 25 was originally conceived Fr

Series 8000 Xpress PSE User Guide

Contents

Download Pdf Manuals

Related Search

Related Contents