Hi Viktor,

I disagree with a couple of points you make:

* Viktor Dukhovni <[email protected]> [141014 10:30]:
> And in particular the "MUST NOT" for RC4 is also inapplicable with
> opportunistic TLS.
>
> [...]
>
> 3.
> SSL Protocol version recommendations.
> 
> Once again with unauthenticated opportunistic TLS even SSL 3.0 or
> TLS 1.0 is (much) better than cleartext.  So the MUST NOT SSL 3.0
> and SHOULD NOT TLS 1.0 are too strong.

So once again we have the whole 'opportunistic' discussion. I
strongly disagree that both should be changed from MUST NOT to a
SHOULD NOT or even something less. The whole point should be to
deprecate SSLv3 and RC4 as soon as possible, no matter what. I do
not think it does matter if the WG sees 'opportunistic' encryption
as sufficient - both have to go. They are a real world security
threat. In particular downgrade attacks on unauthenticated TLS
will enable less-than-optimal security for server and client with
SSLv3 and might enable various attack vectors. For example, the
recent virtual host confusion attack [0] relies on this fact. I'm sure
there are other attacks. If we're talking 'opportunistic' again: as
far as I can remember SSLv3 is succeptible to replay attacks on
anonymous diffie-hellman [1].

As said, I think we should deprecate RC4 and SSLv3 ASAP.

I'm happy that Google does practically the same thing with
deprecating SHA1: once their browser and android devices will issue
a warning, every snakeoil CA needs to act or will loose customers.
Sometimes you have to push to get good security widely deployed.


Aaron

[0] - http://bh.ht.vc/
[1] - https://www.schneier.com/paper-ssl.pdf




Attachment: signature.asc
Description: Digital signature

_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to