Re: [TLS] Products supporting TLS 1.0 & some other high-level questions

[Peter Gutmann <[e-mail]>]

Eric Rescorla <[e-mail]> writes:
>On Sun, Oct 5, 2014 at 7:22 PM, Watson Ladd <[e-mail]> wrote:
>> Is the prefered path
>> -Adoption of TLS 1.3
>> -Adoption of TLS 1.2+session_hash fix
>> -Indefinite support for TLS 1.0 plus multiple, not widely deployed fixes.
>
>As a practical matter, we're likely to get at least the first two and
>probably all three.

You forgot one important word: eventually.  TLS 1.1 + add-ons we've got now.
TLS 1.2 + session hash fix is still a work in progress after six years (eight
if you take it back to the draft versions, which were mostly implementable
even then).  TLS 1.3, which should really be called TLS 2.0 because there are
so many fundamental changes, could take a decade or more (based on the 1.2
timeline) to see widespread deployment.  I'm sure Google will rush it out in
Chrome and bundle it with HTTP 2.0 (at least they're calling that 2.0 rather
than pretending it's still a 1.x version) because it's of benefit to them, but
for non-browser use you're going to see a long tail worthy of the Midgard
Serpent.  So I think you'll end up with a split, "TLS for browsers" (1.3
a.k.a. 2.0) and "TLS for everyone else" (1.1 or maybe 1.2).

Peter.