Re: [TLS] Encryption of TLS 1.3 content type

[Alfredo Pironti <[e-mail]>]

On Fri, Aug 8, 2014 at 3:01 PM, Martin Rex <[e-mail]> wrote:

> Alfredo Pironti wrote:
> >
> > For example, Tor abandoned doing authentication via TLS renegotiation*
> > because, among other reasons, "TLS renegotiation to become rarer and
> rarer
> > in the wild, making our own use stand out more" [2].
> > If TLS could be used to disguise, both in content type and in traffic
> > shape, a renegotiation as normal application data, projects like Tor may
> > benefit from it.
>
> What you may have missed in the discussion: this is not about changing
> the presence, availability and visibility of renegotiation in TLSv1.2,
> TLSv1.1 and TLSv1.0, but exclusively about the visibility of the
> ContentInfo in a future TLSv1.3 _which_has_Renegotiation_removed_.
>

My understanding was that some form of re-keying or client-authentication
may be provided instead. In the latter case, similar reasoning would apply.

If the content type was encrypted, one could use different application data
content type values to multiplex several application streams over one TLS
connection.

Alfredo


>
> -Martin
>