Browse Prior Art Database

Detour Routing

IP.com Disclosure Number: IPCOM000053037D
Original Publication Date: 1981-Aug-01
Included in the Prior Art Database: 2005-Feb-12
Document File: 5 page(s) / 33K

Publishing Venue

IBM

Related People

Bharath-Kumar, K: AUTHOR [+3]

Abstract

A technique is presented to re-direct a single explicit route via an existing transmission group with minimal effect on other routes, despite the system network architecture (SNA) routing constraints. 1. INTRODUCTION

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 32% of the total text.

Page 1 of 5

Detour Routing

A technique is presented to re-direct a single explicit route via an existing transmission group with minimal effect on other routes, despite the system network architecture (SNA) routing constraints. 1. INTRODUCTION

A brief description of explicit routing enhanced with swapping as used in known SNA is first presented. Between every pair of subarea nodes up to eight loopfree explicit routes (ERs) are defined. Thus. a node B will have up to eight ERs defined to every destination node A, wherein these routes are denoted by the pair
A.ERN, where ERN is the "explicit route number" between 0 and 7. The routes are reversible in the sense that given the route A.i defined from B to A, there must be a route B.j defined from A to B which travels along the reverse sequence of transmission groups (TGs). Furthermore, the routing tables used in the nodes are source independent in that the outbound TG for a message arriving at an intermediate node C is determined by its destination subarea and ERN, irrespective of its source. Finally, the routes are labelled with ERNs in a swapping fashion, that is, all messages arriving at intermediate node C on A.i are transmitted on the outbound TG for A.j where j may not equal i.

The fundamental technique of detour routing will now be introduced. The global concepts behind detour routing are first discussed, and then the particulars of detour routing to two specific instances, are covered, namely: the redirection of ERs away from TGs that are operative, and the redirection of ERs away from TGs that have failed. 2. THE NEED FOR DETOUR ROUTING

The general problem addressed is depicted in Fig. 1. A node B has an outgoing route A.1 on TG(1), and B desires to reroute A.1 to TG(2) with minimal global effect. In particular, no other routes should be changed.

After the detour is accomplished, session outage notification will be modified for users of A.1. Also, we must insure that loop freedom is maintained in the resulting routes. Finally, modifications are needed for the RWRI (reset window response indicator) used for virtual route window flow control. 3. THE SOLUTION

To accomplish the desired redirection, each node maintains "detour routing tables" which are devised at SYSGEN time. These tables provide loopfree detours: the entry for node B and route A.1 may not cause any loops.

If a particular explicit route is detoured more than once, then loops may occur even though the tables provide loop-free detours. We introduce a mechanism to prevent multiple detours of a route.

The mechanism to avoid multiple detours is as follows. Each node maintains a bit for each outgoing explicit route which specifies whether the route may be detoured. Also, each node maintains a bit for each explicit route which specifies whether a route may be detoured to it. For example, assume in Fig. 1 that B wanted to switch A.1 to TG(2) and that D swaps A.1 to A.2. B verifies from its bit

1

Page 2 of 5

that A.1 may be detoured an...