Re: [keyassure] Opening issue #21: "Need to specify which crypto

[Phillip Hallam-Baker <[e-mail]>]

On Fri, Mar 4, 2011 at 9:39 AM, Paul Hoffman <[e-mail]> wrote:

> On 3/4/11 5:57 AM, Phillip Hallam-Baker wrote:
>
>> OK, how about this
>>
>> Define code points for
>>
>> SHA2-256
>> SHA2-512
>> SHA3-256 (reserved)
>> SHA3-512 (reserved)
>>
>
> This would only work if we knew today that there will be algorithms with
> those exact names. Those of us following the NIST hash competition have seen
> that there is a good chance that will *not* be the case: NIST is explicitly
> allowing (some would say encouraging) parameters on the functions.


All the finalists are designed to be used as a drop in replacement for SHA2,
which is a mandatory requirement.

But I do prefer using a hash that incorporates random data in the hash
mechanism. It provides a lot of protection against attack, in most cases an
attacker has to achieve a first preimage attack rather than a second.


So while we could certainly reserve a slot now, it is likely that there will
be further discussion on the choice of mode. And I would really like us to
be having one IETF conversation at that point and not ten with different
combinations of what is mostly the same people.

Since we are going to be using this with TLS it is probably going to make
most sense to consider what algorithms TLS is going to use.


-- 
Website: http://hallambaker.com/