Tim - my memory is also that the discussion tapered off without a conclusion.
I do remember getting a summary e-mail from Bob and Brian (ipv6 WG chairs) about "no consensus for changing M/O bits at this point" (but I can't locate that e-mail); however, my memory of the discussion is that there were two interleaved discussions: (1) improve documentation of what was in the RFC 246[012] specs and (2) consider changes (improvements) to spec for M/O bits and DHCPv6. So, I'm unclear about any conclusions... I've included the ipv6 WG chairs and the authors of draft-ietf-ipv6-ra- mo-flags-01.txt in the CC: list; can any of you clarify the resolution of the discussion, and should the discussion appear on the ipv6 or dhc (or both) WG agendas? - Ralph On Mon, 2005-07-25 at 11:45 +0100, Tim Chown wrote: > On Sat, Jul 23, 2005 at 09:38:21PM -0400, Bernie Volz (volz) wrote: > > Also, what's happened with the M/O bit discussion? Has this been > > resolved in the IPv6 WG or are they expecting us to do anything? Should > > that be an agenda item? > > I think it should, at least briefly. > > My recollection is that the discussion fizzled out without a conclusion. > > There is a specifc (newish) draft on the topic: > http://www.ietf.org/internet-drafts/draft-ietf-ipv6-ra-mo-flags-01.txt > > I'm not sure how strong consensus is on the contents of this draft. > > One of the issues was that there are two bits, with four combinations, > one of which is 'meaningless'. But then an administrator should never > use that combination. People still seem to see some 'confusion' in > the semantics. While questions keep hitting the ipv6 list, one can argue > that the M/O text is certainly not as clear/streamlined as it could be. > > Some argued whether a DHC service indication was needed at all. Just let > the client back off. One question was over unecessary use of > bandwidth/battery on low capacity wireless devices - a hint that no DHC > was available would then be useful rather than the client retrying. > > There were some interesting ideas, e.g. that perhaps all stateless DHC > servers could respond to the same message types as stateful - I think the > JOIN people suggested that. Presumably you'd then only need the M bit, > given M and O semantics were being ushered towards the message type. > > I don't like the idea of changing things now without larger deployment > experience, and I recall most people agreed that was lacking. > > Anyway, Ralph/Stig should maybe check that the topic is on the ipv6 WG > agenda and take it from there, since ipv6 is held on Tuesday and dhc is > on Thursday. > > Tim > > _______________________________________________ > dhcwg mailing list > dhcwg@ietf.org > https://www1.ietf.org/mailman/listinfo/dhcwg -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
