As I just said in a separate message, one big question had been raised about rfc2462bis. It was, in my understanding,
whether we need the M/O flags for the "stateful" protocol(s) in the first place. (Actually, different people used different representation related to this issue, but, in my understanding, we can simply summarize the essential part of the question as above). I know this is probably highly controversial, but I tend to agree on removing (or deprecating) these bits. The main reasons have already been shown in the thread starting at the following URL: http://www1.ietf.org/mail-archive/working-groups/ipv6/current/msg02262.html One additional point is that I'm not sure if we can really recycle rfc2462bis as a draft standard with keeping these bits. If we keep them, we'll probably need to find running implementations regarding these bits that are interoperable. (If so) can we really make it soon? As far as I know, no host implementation supports the M flag. Regarding the O flag, I've implemented the host side of the flag, which invokes a DHCPv6 client upon receiving an RA with the O bit. But I've never heard of other implementations. This also means one obvious problem when we remove these bits is actually a non-issue. That is, we do not have to worry about the problem that we may make serious loss of compatibility. Another possible disadvantage is that we'll lose the ability to trigger DHCPv6 (or, if you like a general wording, trigger a stateful protocol). I admit this can be a problem. A rough idea to deal with this is to provide a different, more general way to provide the ability as a separate document. For example, a separate ND option which can trigger a particular protocol (including DHCPv6) for a particular purpose (e.g., to get "additional" configuration information) might make sense to implement the idea. In fact, an extension to ND to re-invoke a protocol for "other information" has already been proposed: draft-vijay-ipv6-icmp-refresh-otherconf-00.txt. I know this is probably too radical to make a concrete proposal at this stage, but if we can really agree on the basic idea of this path, I'd probably do the followings in rfc2462: - remove all references to "stateful" protocols - deprecate the M and O flags, that is, + reserve the fields and clarify that they must not be used (this may affect rfc2461bis, too) + remove all the usages regarding these flags Any comments/suggestions/objections are welcome. Thanks in advance, JINMEI, Tatuya Communication Platform Lab. Corporate R&D Center, Toshiba Corp. [EMAIL PROTECTED] -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
