As a participant, I don't have any particular skin in this argument, so I'll go 
with consensus, which I'm supposed to do as co-chair anyway.  :-)

The one thing not mentioned so far is the fact that there's tell of sites that 
have implemented the H (hash-key) scheme as we currently have it, versus HMAC.  
More specifically, there's one large European ISP that has said they have 
implemented what the -04 version of this draft said.  I don't know if there are 
others, nor do I know if they are tracking this work well enough to know they 
should change if the WG decides to move to HMAC.

We can't argue, though, that this creates an interop concern with the existing 
deployed base, because there's no actual interoperation in the traditional 
client/server sense here; if one party is doing it with H and one is doing it 
with HMAC, they're going to get the same results in the end anyway, minus a few 
possible paths to stealing the redaction key (even though, as we've said, there 
are easier ways to mount that attack) in the HMAC case.

-MSK
_______________________________________________
marf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/marf

Reply via email to