Re: [TLS] The risk of misconfiguration

On Wed, May 7, 2014 at 1:24 PM, Nico Williams <nico@cryptonector.com> wrote:
> On Wed, May 7, 2014 at 2:22 PM, Alyssa Rowan <akr@akr.io> wrote:
>> On 07/05/2014 20:10, Salz, Rich wrote:
>>>> I think that they are insecure _is_ a compelling reason: we seem
>>>> simply to disagree on that point.
>>> No, I think there is disagreement as to whether or not they are
>>> insecure.
>>
>> Because the application layer provides all its own security?
>
> It provides its own authentication, but not its own security layer.
>
>> Then why would it need TLS; decoration for middleboxes¹?
>
> Because it's simpler this way.
>
> SASL used to provide a "security layer".  It's now deprecated because
> it only led to problems:
>
> a) double encryption in some cases (since SASL apps generally also use TLS)
> b) extra code in the application and the SASL mechanism, all variously
> duplicating TLS functionality, not always well done.
>
> With channel binding both of those go away.

But you get a new issue: channel binding doesn't work with all
ciphersuites because of the DHE validation issue. I think SASL is
important enough to support anonymous ECDH, and maybe anonymous DH if
we can fix it. However, using it without authenticating endpoints
isn't getting anything.

Sincerely,
Watson Ladd
>
> Nico
> --
>
> _______________________________________________
> TLS mailing list
> TLS@ietf.org
> https://www.ietf.org/mailman/listinfo/tls

-- 
"Those who would give up Essential Liberty to purchase a little
Temporary Safety deserve neither  Liberty nor Safety."
-- Benjamin Franklin