santuario-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Scott Cantor" <canto...@osu.edu>
Subject RE: C++ lib support for SHA-256, etc.?
Date Thu, 17 Feb 2005 15:57:43 GMT
> Interesting point.  If one of the required algorithms is really broken,
> it may be difficult to reach consensus on what to use in its place.
> Implementation A may support algorithm 1 but not algorithm 2, while
> implementation B supports 2 but not 1, and the two no longer
> interoperate.

Right, that's my concern. Ideally one could quickly patch the applications
to switch algorithms (if they made the mistake of not simply allowing the
deployer to pick at runtime, <cough>) but only if there're alternatives that
are supported.

> Is this the sort of thing you're concerned about?  If so, I'm not sure
> what can be done about it.  The algorithms that are available to be
> included in specifications today may be cracked tomorrow.  Maybe what's
> needed is a formal process for incorporating new required algorithms
> into the specifications that is less cumbersome than the normal
> specification process.  (Of course, that process would have to be
> specified...)

I think a starting point is not letting the vendors slide by during
conformance discussions by insisting that only the bare minimum be MTI.
That's not going to cut it. Right now, anything but RSA-SHA1 is essentially
worthless if you're dealing in interoperable signatures (excluding DSA, I
suppose).

I'm a spec implementer myself, so I understand the impulse.

> A thread on the W3C Dsig mailing list has begun to address the question
> of alternate algorithms (see
> http://lists.w3.org/Archives/Public/w3c-ietf-xmldsig/2005JanMa
> r/0030.html), but has not gotten very far (yet).  Maybe you should chime
> in.

Thanks for the pointer.

-- Scott


Mime
View raw message