OSPF failure reconvergence through SRG inference and prediction of link-state advertisements
Bart Puype1, Dimitri Papadimitriou2, Goutam Das1,
Didier Colle1, Mario Pickavet1 and Piet Demeester1,
1: Ghent University — IBBT
2: Alcatel-Lucent Bell
Abstract The ECODE FP7 project researches cognitive routing functions in future networks. We demonstrate machine learning augmented OSPF routing which infers SRLGs from network failure history. Inferred SRLGs are used to improve OSPF convergence and recovery times during subsequent (multiple) network failures.
Demonstration video : capture.avi (2:10) 66 MB (mpeg4)
