> > Keith Moore wrote: > > sorry, I thought you were using "site-local" in a broader > > sense than that which is in the these documents. > > What kind of broader sense?
people have been talking about globally unique site locals. > I am sympathetic to ambiguity being a pain in the kazoo, but it does > guarantee to some extend that SLs are not publicly routable. I'm all for > removing ambiguity, but not at the cost of removing the lack of > routability at the same time. lack of routability is not entirely a feature. it's more like a mixed bag. constraining the routing table might be a laudable goal. otoh it can't be justified from a security standpoint. > Since you agree that FEC0::/10 should not be used for publicly routable > addresses, what issues do you have with having FEC0::/10 used as > globally unique, not globally routable, site-local addresses (which > could be achieved now) *and* another future block for aggregatable, > globally unique, globally routable addresses (which will take more time > to get)? imho we have too many different kinds of addresses already, and we're putting way too much burden on the idea that hosts should be able to select which addresses are best for their purposes and the network conditions. so while I agree in principle that we can't use FEC0://10 for GUPIs, I hope that in practice FEC0://10 is mostly used for isolated networks. Keith -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
