Hi,

On Wed, 6 Nov 2013, Randy Bush wrote:

The only reason local, validated ROA databases can be different is
because of the update interval of the validation software that
operators have running, i.e. I may have just fetched an update with
new ROAs that you don’t have yet.

network partition affecting fetches of some repositories by some client
caches.  some caches hung for ghu knows how long dealing with apnic's
deviant repository structure.  ...

Okey, so you want to build routing policies based on a non-steady state? I'd like to keep things simple and this surely doesn't make live easy. So I'd be building my routing policies on a steady state (and ROA databases would, with time, converge to a steady state).

Any thoughts on the reasoning I made for not prefering valids over unknowns?

Jac


--
Jac Kloots
Network Services
SURFnet bv
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to