Re: [TLS] Security concerns around co-locating TLS and non-secure on same port (WGLC: draft-ietf-tsvwg-iana-ports-08)

[Nicolas Williams <[e-mail]>]

On Thu, Nov 11, 2010 at 05:41:15PM +0100, t.petch wrote:
> > Philosophy is fine, but the original poster proposed something concrete
> > (that we drop StartTLS and just always use raw TLS); we should have an
> > answer to that.  Mine is that the arguments against StartTLS are weak,
> > and the arguments for always using raw TLS are also weak.  I'm
> > unconvinced by the OP's and others' arguments against StartTLS.
> 
> To return to (what I see as) the main purpose of the thread, I too
> think that StartTLS is a good, if not an excellent idea; I see no 
> difference in the vulnerabilities (although my cryptanalysis is weak).

There's no need for cryptanalysis w.r.t. what happens before TLS starts:
it must all be in the clear (relative to the application; IPsec could be
involved, of course).

> Sadly, whenever I have been involved in a WG developing App-X 
> over TLS, my record of convincing the rest of the WG is 100% failure.

I think it's fine for _some_ applications to require TLS at all times.
For most applications I'd prefer StartTLS, but that's due to my port-
preserving bias.  Looks like we need guidance from the IETF here:

> Also, this decision, StartTLS or separate ports, tends to be taken 
> early in the development cycle, while the request to IANA will only materialise
> five years or so later, so back pressure from IANA or the IESG will then meet
> an immovable mountain.  Really, any advice to a WG needs to put in front
> of them about the time the BOF is approved, which I don't see this I-D 
> having any material impact on!

We should probably have a BCP indicating which method is preferred, so
we can flog it at WGs early on.

Nico
--