[c-nsp] Segment Routing (SR)

[Adam Vitkovsky]

On Saturday, August 24, 2013 6:16 PM Mark Tinka
wrote:

> I don't dispute that BGP as a potential IGP will scale 
> (look at iBGP today), but will it be quick? 
> That's my concern.

Well my point is it doesn't really need to be quick. 
With pre-programed loop-free backup path at each hop I don't really care
about how fast the change in topology is propagated throughout the backbone.


And with great innovation from Ahmed Bashandy, that's promised to be
incorporated into SR, even the need for IGP to signal the failure of egress
PE to ingress PE as fast as possible so that BGP can invoke PIC-core is
rendered unnecessary. 
So therefore I believe there are no more reasons for IGP to be fast in
topology change propagation. 

adam