On balance, I prefer ducking the issue in the draft we are discussing. If I get a /48 I have 16 bits for subnet addressing and I'm happy, so why worry about the acronym SLA? The RIRs have accepted the point (but I still want to see an informational reference to 3177 to ensure the paper trail is complete).
Brian Michel Py wrote: > > Erik / ipv6 folk, > > > Erik Nordmark wrote: > > RFC 2374 contained an IPv6 allocation structure that > > included TLA (Top Level Aggregator) and NLA (Next > > Level Aggregator) which is formally made historic by > > this document. > > The TLA/NLA scheme has been replaced by an coordinated > > allocation policy defined by the Regional Internet > > Registries [REF]. Part of the motivations for obsoleting > > the TLA/NLA structure were technical, for instance there > > was concerns that it was not the technically best > > approach at this point in time on IPv6 deployment. > > Another part of the motivation was that the issues of > > how the IPv6 address space is managed is much more > > related to policy and to the the stewardship of the IP > > address space and routing table size that the RIRs have > > been managing for IPv4. It is likely that the RIRs > > policy will evolve as IPv6 deployment proceeds. > > The IETF has provided technical input to the RIRs (for > > example [RFC 3177]) that has been taken into account > > when defining the policy. > > I was re-reading your text, and I have additional comments. > > [disclaimer: I have stated before that it was fine by me, and this still > holds. If consensus is reached on the text you proposed the doc should > be shipped without further delays.] > > That being said, I have contradictory / ambiguous feelings about the > omission of "SLA". Here's the contradiction: > > - On one side, since you explicitly kill TLA and NLA but not SLA, it is > permitted to think that SLA is not completely dead, which suits me fine > since my position is that although moving it to policy was a good idea, > killing the notion of site boundary was not. > > - On the other side, if SLA is not dead it's not alive either, which > makes it a zombie I guess. This is not good, and one of these nights, > when the moon is full, it will rise from the grave like in Michael > Jackson's "thriller" and come haunt us. > > Therefore, we have three options for SLA: > 1. Don't change the text and keep it a zombie. > 2. Kill it (which I don't like). > 3. Spare it; the idea being that what we want to do is move it to policy > but don't kill the notion that a site boundary exists. This is what RIRs > call a "soft" or "semi-hard" boundary. > > Comments welcome. > > Michel. -------------------------------------------------------------------- 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] --------------------------------------------------------------------
