Re: [TLS] New cached-info draft 09 posted

[Stefan Santesson <[e-mail]>]

Michael,

You are absolutely right and I was totally wrong.

In order to preserve the security properties of TLS 1.1 and below we would
have to use a combo which is not compatible with the current syntax.

Using the PRF seems however to have a similar problem. You would have to
indicate which PRF you are using, and if TLS 1.2 PRF is used, what hash
algorithm to use. That doesn't feel trivial either.

/Stefan




On 10-07-13 7:40 PM, "Michael D'Errico" <[e-mail]> wrote:

> Stefan Santesson wrote:
>> 
>> We had a very long debate about this and we finally reached an agreement.
>> Can we stick with it or do we have to redesign this over and over again?
> 
> I'm not sure if you were disagreeing with me, so I'll clarify.  My point
> was that this struct:
> 
>        struct {
>             CachedInformationType type;
>             HashAlgorithm hash;
>             opaque hash_value<1..255>;
>        } CachedObject;
> 
> uses HashAlgorithm from RFC 5246.  There is no value for the MD5/SHA-1
> combo hash used in TLS 1.0 and 1.1, so we need to pick something that
> has an identifier.
> 
> Mike