[c-nsp] is RPF strict mode common?

[Mark Tinka]

On Thursday 08 May 2008, Adam Greene wrote:

> Trying to control bandwidth between my (2) upstream
> Internet providers, Global Crossing (20Mbps) and Savvis
> (50Mbps). I currently receive full routes from both, and
> the smaller Global Crossing link is maxed out, inbound.
>
> The obvious solution to me will be to prepend my route
> announcements to Global Crossing. However, one question:
> there is a good chance that some of my traffic will flow
> out through Savvis and in through Global Crossing (in
> fact, that's almost certainly happening right now). Will
> this kind of asymmetrical traffic run into issues with
> other ISPs that deploy RPF in strict mode? Are there many
> ISPs out there that do this? It seems that so much
> traffic on the Internet must be asymmetrical, any ISPs
> running RPF in strict mode must be doing so in a way that
> will not break traffic that's asymmetrical because of
> other ISPs' standard routing policies. IF they do, then
> they would be causing dead spots for their own customers
> ... do you think that's a valid assumption?

We use strict mode for customer connections - specifically, 
customers that do not multihome, either to ourselves or 
another provider.

We "generally" use loose mode for peering (upstreams, public 
peering, private peering, e.t.c.).

We don't use uRPF in our core (BGP-free core and all 
that...).

Cheers,

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 832 bytes
Desc: This is a digitally signed message part.
Url : https://puck.nether.net/pipermail/cisco-nsp/attachments/20080508/9b5f51ce/attachment.bin