A self-healing protocol for ATM networks based on multiple backup VPs

Asynchronous transfer mode (ATM) networks permit the establishment of backup virtual paths (VPs) with zero bandwidth. This feature makes ATM layer based self-healing mechanisms particularly interesting as faster restoration can be realized by capturing the necessary bandwidth on the backup VP after a failure. With the aim of further increasing the restoration speed, this paper studies a new multiple backup VPs (MBV) based protocol. Moreover, a new approach for capturing bandwidth on a pre-established backup VP is presented, where, after a failure, nodes on the backup VP start capturing the bandwidth in a quasi-parallel way as opposed to the traditional step-by-step way. Depending upon the length or properties of a working VP, the MB protocol allows one or more backup VPs. This approach helps holding the restoration information in a reduced area, leading to a faster recovery. Results show that the MBV protocol considerably improves the speed of restoration from a failure as compared to a better-known end-to-end VP based protocol. Moreover, the proposed way of bandwidth capturing on the backup VP also offers better speed. The MBV protocol can employ multiple backup VPs in an overlapped or non-overlapped way. Operation of both approaches is described and performances are evaluated in terms of two criteria, namely restoration ratio and the amount of additional transmission capacity required to be employed