> If that is the case, having a set of policy objects expressing AS > relationship should do the same thing and more with less overhead?
real policy is per prefix, customer, peer, and things disgustingly more complex, with complications of backdoor relationships, ibgp policies to implement regionalization, and a jillion things you would not want to tell your children. aside from not wanting to publish much of that, even if it was publishable, how would you differentiate 'the announcement really took this path' from 'it is possible that the announcemnt took this path?' the latter being a half-assed probabilistic assurance. randy _______________________________________________ sidr mailing list sidr@ietf.org https://www.ietf.org/mailman/listinfo/sidr