On Tue, May 16, 2017 at 12:04:57PM -0700, Eric Rescorla wrote: > On Tue, May 16, 2017 at 11:30 AM, Nico Williams <[email protected]> > wrote: > > On Tue, May 16, 2017 at 01:43:32PM -0400, Richard Barnes wrote: > > > As has been pointed out elsewhere, other key changes are signaled with a > > > handshake message (KeyUpdate), so using a handshake message seems more > > > natural from a protocol point of view. > > > > And as long as the record type goes in the clear, sending these sorts of > > messages all with the same record type (handshake) seems best from a > > traffic analysis p.o.v. > > Actually at this point in the handshake the record type is encrypted.
Oh good. _______________________________________________ TLS mailing list [email protected] https://www.ietf.org/mailman/listinfo/tls
