Re: 9.2.2 Cipher fallback and FF<->Jetty interop problem

[Eric Rescorla <[e-mail]>]

On Mon, Sep 22, 2014 at 9:41 AM, Jason Greene <[e-mail]>
wrote:

>
> On Sep 22, 2014, at 11:18 AM, Eric Rescorla <[e-mail]> wrote:
>
> >
> > I don't actually think this is that important an issue either. As I
> understood the discussion
> > in Zurich, the new TLS limitations were directed towards pulling users
> of HTTP2 towards
> > modern algorithms. However, algorithms which have serious weaknesses
> should probably
> > be deprecated in all versions of HTTP (as with
> https://tools.ietf.org/html/draft-ietf-tls-prohibiting-rc4-00)
> >
> > Say we decided that in future we preferred Aero (
> https://tools.ietf.org/html/draft-mcgrew-aero-01)
> > to AEAD constructions. That seems like something we could roll out in
> HTTP3 but wouldn't
> > be appropriate to retroactively apply to TLS 1.2 unless there was
> something seriously wrong
> > with AEAD (and then see above).
>
> I think this hypothetical actually counters your point. Every rev of the
> HTTP spec introduces interop cost, therefore having to rev the protocol
> just because TLS needs to rev is unnecessary cost.


I think we're talking past each other here. There are two major cases:

- We're kind of sad that people use algorithm X and we wish they would
   do something more modern.
- There is something seriously wrong with algorithm X and people need
  transition off it pronto.

In the former case, we have pretty limited options, since it's probably not
worth breaking interop over. So, we can do nothing or we can gradually
tell people to upgrade at preexisting protocol upgrade points. I.e., we
wouldn't roll out HTTP3 to do this, we'd just do it when we were already
rolling out HTTP3 (the same way as 9.2.2 is now). in the second case,
we would want to adjust all versions of HTTP so no new rev would be
required.

-Ekr


-Ekr