[c-nsp] Resilience in order of few hundreds of milliseconds

[Mikael Abrahamsson]

On Tue, 6 Mar 2007, Alaerte.Vidali at nokia.com wrote:

> Looking for alternatives of fast recovery of MPLS under failure without 
> using FRR. Any input appreciated.

Run your MPLS over L2TPv3 and don't run MPLS natively in the core. Tune 
IGP agressively and you should be able to converge in 100-300ms.

You of course need a core link technology that supports hardware failure 
signalling (POS for instance) to detect failure to trigger reroute.

PS. This goes under the "any input" you requested. It's doable, but it 
requires a lot of changes in your network, most likely.

-- 
Mikael Abrahamsson    email: swmike at swm.pp.se