Re: [Cfrg] Request For Opinions

Alex Alten  wrote:
>At 06:05 PM 5/10/2003 +0000, David Wagner wrote:
>>I don't think you've thought about this carefully enough.  Use public-key
>>crypto to exchange a shared AES key, and then you can re-key using that
>>AES key just as you would if you weren't using public-key crypto.
>
>OK, I agree this is a scenario where RSA is used to exchange a single AES
>over a long session of packets is fine.  But if I'm doing single packets, and
>changing AES keys a lot, it doesn't work very well.

So don't do that, then.  No one said you need to use RSA on every packet.
Come on.  This is getting really silly!

Patient: "Doctor, doctor, it hurts when I use RSA on every packet."
Doctor:  "Don't do that, then."

>That's not what I mean't.  SSL has both high PK and hash costs.  Most of
>my response was discussing the PK performance penalty limiting high-
>performance design choices.  David at least you can admit to the
>reasonableness of that statement.  Right?

That's ridiculous.  The public-key in SSL has nothing to do with whether
its hash function or MAC primitive is fast or slow.  Again, I don't think
you've thought about this carefully enough.
_______________________________________________
Cfrg mailing list
Cfrg@ietf.org
https://www1.ietf.org/mailman/listinfo/cfrg