I agree completely with Carsten on this. It is NOT up to the WG chairs to determine direction or consensus. It is up to the WG.
I am in favor of ND-09. We should have some implementation experience shortly but it does appear to address the stated issues to ND-08 and appears to be a solid basis for 6LowPAN ND. Don -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Carsten Bormann Sent: Monday, May 10, 2010 5:13 AM To: [email protected] Subject: [6lowpan] Question of the week: ND-09 status >> I've not seen a consensus to rewrite ND from 08 to what it is >> today. I've seen Geoff at the mike in favor, and you saying don't worry >> I'll work with the author of ND simple and sort that out. Hardly a >> consensus to me. > > WG direction is an issue for the chairs... Well, no, this is an issue for the WG. There was some vocal opposition to the complexity of ND-08. Since I was a co-author of ND-08, I wasn't in a position to judge consensus as a WG chair, but if I had been, I would have said very, very rough consensus at best. I didn't co-author ND-09. I do think personally that ND-09 is in reasonable shape, and I haven't heard strong technical arguments against it. With the exception of Pascal, all ND-08 authors seem to like it (even though each of us probably would do some little thing in a different way). But, very importantly, this is not a paste-over-the-disagreements compromise solution: It is a solid redesign based on some evolved assumptions on what the base 6LoWPAN-ND protocol has to do and what not. Now my questions to the WG: 1) Are these evolved assumptions correct? The most important change is that the assumption is now that we can fully rely*) on the uniqueness of EUI-64 identifiers. Another assumption is that staying as close to the formats of RFC 4861 as possible is a good thing. 2) Is the protocol defined in ND-09 successfully solving the problem? If the above assumptions are correct, is ND-09 the right protocol? Does it work? If you think that additional functions such as backbone router, or possibly using DHCP servers for certain configuration functions, are needed: can these be added as separate documents**), with hosts implementing just ND-09 still working correctly? 3) Is the document that is ND-09 good enough? Does it define the protocol unambiguously? Is the terminology acceptable? I would like the WG to make some progress on these questions this week. Please speak up. Gruesse, Carsten *) [From my personal point of view, this is mainly burden-shifting to the commissioning/security process, but it is not unreasonable to do so.] **) [Yes, it would be nice to have these documents, to make sure we understand fully how they fit together with ND. But I don't think we have to wait for all of them.] _______________________________________________ 6lowpan mailing list [email protected] https://www.ietf.org/mailman/listinfo/6lowpan _______________________________________________ 6lowpan mailing list [email protected] https://www.ietf.org/mailman/listinfo/6lowpan
