On Thu, 2005-06-23 at 15:15 +0100, Mark K. Thompson wrote: > On 23 Jun 2005, at 15:03, Vlad Yasevich wrote: > > > > Wouldn't an update to a policy table be enough to resolve the problem? > > It's more than one update, I think:
My point was more that an update (or updates) to the policy table would solve this. I guess the number of updates depends on how the policy is implemented. You could always prefer global over ULA for multicast, and will fall back to ULAs only if all global prefixes are deprecated. This is really and end-user decision. -vlad > > If the candidate destination address is from a fc00::/7 prefix > in the set that are either local to this network or a prefix at a > network with whom I have routing agreements, then prefer a local ULA > source address; otherwise if the candidate destination address is > from any other fc00::/7 prefix, never prefer our ULA source address > > Perhaps the second clause is not necessary where the first rules bump > local ULAs up the preferred source address list (one rule per > routable fc00::/7 prefix) > > I can see a straw poll of 3484 implementations on the horizon... > > Mark/ > > -- > Dr Mark K. Thompson > Electronics and Computer Science > a School of the University of Southampton, UK > > > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- > -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
