Le 2013-01-30 à 10:04, Ole Troan <[email protected]> a écrit : > Remi, > >>> can we please hold back a little and let other people in the working group >>> voice their opinion? >> >> >> To make progress, I feel it should be acceptable to continue to: >> - ask for clarifications on some points that IMHO remain obscure (e.g. the >> purpose of overwriting "the most significant parts of the interface >> identifier" for CE IPv6 prefixes of lengths > 64, especially since it >> conflicts with RFC 6052). > > I don't quite agree that it conflicts, but looking back at comments from > several people it might be best to remove the 6052 reference.
Removing the reference does eliminate the conflict. What remains unknown is the purpose of IPv4-address and PSID fields if they may be partially overwritten, e.g. with IPv6 prefixes of 96 or 120 bits. >> - react to statements or assumptions that are AFAIK inaccurate (e.g. that, >> with MAP-E IID formats as specified, maintenance could determine PSID values >> from IIDs alone.) > > agree, Good. > but you can get a 'fair' idea. given that both the IPv4 address and PSID are > fixed length fields at fixed offsets. 'Fair' idea is unclear but yes, if the PSID length is known to be fixed, one can find the PSID (only in that case). > I'm defending status quo largely because that gives me the least amount of > work. ;) > > please convince the working group of your scheme, and provide the text, and > I'll put that in the document. > so far I have seen more support for the existing scheme than the "::1" though. Provided the IID "overwriting" is clarified (see above), I agree that whatever the majority prefers will become the standard, even if unnecessarily redundant. My contribution is simply to suggest an application of the KISS principle. Regards, RD > > cheers, > Ole _______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
