RE: On 5.5 key agreement

Merlin replied to me, 

> The KM function is a standard mechanism used by ANSI 9.42 (and others)
> and from there by IETF documents (e.g., RFC 2631) <skip>

I did not mean to protest against the particular KM function. My complaint
was that the KM function should not be `burned in` the spec as there may be
reasons, such as a possible future crypto weakness in the particular KM
construction, for people to prefer another KM function. Furthermore I noted
that the KM function should be part of the AgreementMethod. Indeed RFC 2631
defines a particular key agreement method (Diffie-Hellman of course) and
therefore the fact he defines the KM construction there is exactly in line
with my suggestion. 

So...
> >My proposed solution, as I indicated above, is simply to 
> have this aspect
> >covered by the AgreementMethod and removed from the spec. 

Best, Amir Herzberg 

Received on Monday, 3 September 2001 03:46:58 UTC