The Design of Survivable ATM Networks

Abstract

It is essential that broadband communication networks are survivable. Failure of links or nodes could have disastrous consequences in a very high speed network carrying large volumes of data, for the user, for the service provider and for the network operator. Hence, restoration mechanisms are required which automatically reconfigure routes to avoid the location of a failure. In this work, a backbone ATM mesh network is considered whereby Virtual Paths are cross-connected at special ATM switching nodes. Initially, the means of performing restoration given a certain spare capacity distribution which does not guarantee comprehensive recovery, is examined. This encompasses the preparation required in advance of a failure, and the actual execution of restoration. A network design scheme is subsequently described which enables provisioning of spare capacity for complete restoration from particular failures: the expected Virtual Path traffic demands and routing must be known a priori.