haas@utah-cs.UUCP (Walt Haas) (04/17/84)
I am quoting from a paper entitled /An X.75 Based Network Architecture/ by Donald F. Weir, John B. Holmblad and Amy C. Rothberg which appeared in the /Proceedings of the Fifth International Conference on Computer Communications/, October 1980. >From the abstract: "...Recommendation X.75 was adopted by the CCITT as the interface protocol for interconnecting public data networks. GTE Telenet has used X.75 for its intranetwork protocol as well as for inter- network connection." >From page 742: "Just as X.25 consists of three layers, X.75 also consists of three layers. The Physical Level is defined to accomodate higher speed circuits (e.g. 50 kbps)., using V.35. The X.75 Frame Level is the same as that of X.25 except extended numbering is employed, allowing for the more efficient use of satellite circuits. The X.75 Packet Level differs from X.25 in two ways - the call setup packets contain an additional utility field and clearing/reset causes are passed transparently from network to network. The utility field allows for the signalling of transit network identification, call identification, flow control parameter negotiation, etc." >From page 747: "In the unlikely event that a temporary routing loop or 'infinite' path occurs due to dynamic changes in topology or incorrect configuration information from the NCC, the routing process must detect either occurence and initiate recovery action. This is accomplished by main- taining a hop count and a "node visited" list in the utility field of the X.75 Call Request packet. Whenever a Call Request is received by a node, the hop count in the packet is checked against a network wide maximum and if it is exceeded, the call is cleared back to the previous node. Otherwise, the hop count is incremented and the "node visited" list is examined. If the node finds its node identifier in the list, the call is cleared back to the previous node. Otherwise the node adds its ID to the list and attempts to route the call. In either failure case, the previous node will a ttempt to find an alternate route." Regards -- Walt Haas