On 11/11/16 8:12 AM, Nick Hilliard wrote:
> Sriram, Kotikalapudi (Fed) wrote:
>>> right, ok - I misunderstood.  So you're suggesting that the control
>>> plane correlates asns to interfaces and does something like
>>> creating a higher cost alternative path out each candidate source
>>> interface (based on ASN, as determined in the control plane) to
>>> allow the urpf mechanism handle this using its normal lookup
>>> method?
>>
>> Yes, that is correct.
> 
> I can't help feeling that the use case looks a bit contrived.  The sort
> of situations that could benefit from urpf that I've seen on production
> networks don't generally involve inconsistent prefix sets being
> announced via different connectivity providers, where there is a
> requirement for urpf away from the leaf network.
> 
> If it's any help, the most usual situation where urpf would be handy but
> can't be done, is where you have an interface facing an upstream and you
> want to receive prefixes (and send traffic) via that interface but not
> announce anything (or receive traffic).

The case where routine maintance creates temporary blackholes is
basically ubiquitious were you attempting to use urp strict.

> Nick
> 
> _______________________________________________
> GROW mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/grow
> 


Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow

Reply via email to