On 09/10/2013 09:12 AM, t.p. wrote:
It is a shame that this opportunity was not taken to highlight the need
for authentication. Having a totally secure channel with perfect
encryption is of little value if the other end of the channel is a
hostile power.
True. But if strong authentication at Internet scale is so
hard that people fall back to cleartext then that's worse.
Strong authentication can also in some cases expose identifiers
where you wouldn't otherwise need to, which is not the best
thing from a privacy perspective.
So for at least some of what's recently reported, it seems
to me that there is value in exploring whether opportunistic
encryption is worthwhile, maybe for cases where we don't yet
have strong authentication schemes that are privacy
friendly and that are deployable at Internet scale.
But yes, we also need to worry about strong authentication
and making that easier/better. I'd be happy to see folks
working on this from both approaches - making strong
authentication easier/better but also taking the approach
of seeing whether and when opportunistic encryption adds
value. I would not be happy if we dive into either one while
ignoring the other.
S.
RFC3365, which you cite, gets in right (of course!). It lists three
requirements and top of the list - Authentication service. It may of
course be that the author was only putting the requirements in
alphabetic order but whatever the reason, the emphasis is appropriate.
Tom Petch
----- Original Message -----
From: "IETF Chair" <chair(_at_)ietf(_dot_)org>
To: <ietf(_at_)ietf(_dot_)org>; <ietf-announce(_at_)ietf(_dot_)org>
Sent: Sunday, September 08, 2013 10:53 PM
Here are some thoughts on reports related to wide-spread monitoring and
potential impacts on Internet standards, from me and Stephen Farrell:
http://www.ietf.org/blog/2013/09/security-and-pervasive-monitoring/
Comments appreciated, as always.
Jari & Stephen