[Nets-seminars] Talk GS/302 Friday 8th May 4pm

[Richard G. Clegg]

Don't forget tomorrow's talk by our own Suksant Sae Lor (Hui) on 
"Achieving Full Fast Failure Recovery in Hop-by-Hop Routing".  See you 
there.

Abstract

A novel approach for network routing resilience called Full
Fast Failure Recovery (FFFR) is proposed. It is designed to achieve full
fault tolerance against single link failures while employing minimal
overhead in the routing tables and providing a loop-free environment.
The FFFR algorithm is a modification of the usual shortest path
algorithm used in hop-by-hop routing protocols such as OSPF and IS-IS.
Two routing tables are constructed by using FFFR to perform normal
routing and achieve seamless rerouting in failure scenarios. The routing
tables are obtained by recalculating the link weights based on the
underlying network topology. When a failure is detected, the router
marks the packets in order to allow packet delivery via the alternate
path. FFFR guarantees 100% repair coverage for any recoverable single
link failure. It does not significantly change the way IP routing works
(it does not, for example, employ any form of source routing) and has
minimal impact on routers’ memory since only one extra routing table
entry per destination is required.


-- 
Richard G. Clegg,
Dept of Elec. Eng.,
University College London
http://www.richardclegg.org/