Hi, Ole, On 02/07/2013 09:28 AM, Ole Troan wrote: > Fernando, > >> It's been a month since I received your last message regarding this I-D. >> >> Can we ship this document now? >> >> P.S.: This latest version addresses the feedback received from Tassos >> (I've double-checked with him that his comments have been addressed). > > it would be good if you could summarise the changes from the last call.
I will. > I read through the document again, and my only remaining concern is that the > document doesn't specify the > hash function to be used. > > is there usefulness in having stable interface-ids that are generated the > same across implementations? > and are predictable given knowledge of the "secret key". I would think so. I'm not saying there wouldn't be any value, however: 1) That means all implementations should employ the same hash function -- however, e.g. an embedded device might be more inclined to use MD5 than a more expensive function (that could be used by, say, a desktop computer). 2) Mandating a hash function makes us stuck with a single hash-function forever. What if whatever function we adopt becomes deprecated? (we would update the std, and now have a set of boxes that behave one way, are another set that behaves in another). 3) One of the paramters to the hash function is the Interface Index. Different implementations use different values for the interface index (for instance, the RFCs do not speficify how an interface index should be constructed) -- and, besides, there's no requirement whatsoever that even if two implementations use, say, small numbers for interface index, they use exactly the same values, for the same interfaces. Put another way, it seems to me that what you describe would not be that easy to enforce. Thoughts? Thanks, -- Fernando Gont SI6 Networks e-mail: [email protected] PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492 -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
