On Tue, 7 Feb 2006, Maurice Massar wrote:
[snip]
Thank you for your comments. Since I always take certificate warnings
seriously, I tend to forget that many users just ignore these warnings.
I wish that Kerberos was better deployed (and much easier to configure on
Windows!). Sadly, many systems are set up with a fallback to plaintext
passwords, which IMHO largely defeats the benefit of Kerberos.
Here's another issue that comes up with the UW IMAP client code. Unless
explicitly told not to, it negotiates TLS automatically on all non-SSL
links before it even thinks about authentication.
The story gets worse. Suppose, as site policy, all sessions must have
both integrity and privacy protection; and both clients and servers must
enforce this policy. This means that the server must refuse to
authenticate a session that does not set the security layer to 4 (I'm
assuming that privacy protection implies integrity protection); but only
if SSL/TLS is not in effect. Since the client also insists upon both
integrity and privacy protection, it must abort a GSSAPI negotiation that
does not offer it.
This makes for a very complex initial authentication.
Also, isn't Kerberos vulnerable to a MITM attack on the client->KDC path?
Although Windows insists upon a negotiated client password (which has
other problems), MIT Kerberos does not.
-- Mark --
http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.
_______________________________________________
Imap-uw mailing list
[email protected]
https://mailman1.u.washington.edu/mailman/listinfo/imap-uw
