Re: [TLS] Working Group Last Call for draft-ietf-tls-downgrade-scsv-00

[Florian Weimer <[e-mail]>]

On 10/15/2014 02:43 PM, Salz, Rich wrote:
>> So the number I want to see is: How many servers are there that are not
>> capable of correctly negotiating the SSL/TLS versions they support?
>
> That's not the question that SCSV is trying to address.
>
> The question is:  how many clients end up using SSLv3 when the could use something better?

But the new SCSV will make things worse for accidental protocol 
downgrades (which happen quite frequently with Firefox, it seems).  What 
is now a protocol downgrade will turn into a hard connection failure.

Will the new SCSV enable additional clients to switch on more recent TLS 
versions by default?  I don't think so—you'd still need to code the 
fallback logic separately and outside of the TLS implementation (e.g., 
to support things like STARTTLS, which the TLS implementation typically 
doesn't know about).

-- 
Florian Weimer / Red Hat Product Security