Le 2013-02-07 à 18:02, Doug Barton <do...@dougbarton.us> a écrit :
> Rémi, > > I think you've misunderstood me. Please see below. > > On 02/06/2013 02:01 AM, Rémi Després wrote: ... >> Fair enough. My answer is that no software that I know makes >> determination based on the u bit. > > Thanks, that matches my (admittedly imperfect) understanding as well. > >> My point is that it isn't sufficient for the IETF u bit to have no >> meaning. (a) It has what is specified in RFC 4291. (b) No software I >> know creates IIDs having u = 1 without an IEEE-based OUI. > > So my new question is, why do you care? See below. > > To amplify slightly, I am ambivalent about un-defining the u and g bits. > Generally such a thing is difficult to do, since it's nearly impossible to > get a complete survey of all software everywhere that may use those bits, and > determine if it is safe to stop to just stop caring about them. Since it is at least risky, one would need a good reason to take that risk. > OTOH, I am not a fan of magic bits in general, and I have never seen an > adequate explanation of why an implementor should care about these 2. So I'm > sympathetic to the argument about un-defining them, but you seem to be > vehemently opposed to it, so I'm trying to understand your objection. "It's > important because it's important!" doesn't help me. :) Let me clarify where I stand: (a) Whether changing RFC4291 or not is not the issue for which I work so hard these days. I do think it isn't necessary and have argued in this direction, but my point is that I see a real benefit (point (b)) in using RFC 4291 as is. No modification should break anything, including this benefit. (b) The benefit comes from the following, i.e. one of the 4rd objectives: - We want to statelessly establish automatic tunnels for residual IPv4 across IPv6-only domains. - We don't want, when doing so, to impose any renumbering of IPv6 link and/or host. - This is possible ONLY IF we use, for each CE tunnel endpoint, an IPv6 address that no host in the CE site may be using. - Fortunately this happens to be feasible, with IPv6 as is: in a site that is delegated a prefix up to /64, no host that has a global unicast address conforming to RFC4291 may have an IID in which u=g=1. - By reserving a small subset of this unused IID space, we reach our objective. Hope it clarifies. Regards, RD > > Doug > > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
