string-to-key in Windows Server 2003

[Klyne, Simon]

I believe that W2K3 is switching from UDP to TCP earlier than W2K.

We had an issue with Solaris 8/9 working against W2K but not W2K3 (principle
lengths)
The response we had from Microsoft was that W2K switches from UDP to TCP at
2000 bytes and W2K3 switches at 1500 bytes.  We were able to demonstrate
working TCP functionality to W2K3 by using heimdal rather than MIT kerberos.

Simon


-----Original Message-----
From: Kevin Coffman [mailto:kwc at citi.umich.edu]
Sent: Thursday, June 26, 2003 23:04
To: 'Ben Cox'; kerberos at mit.edu
Subject: RE: string-to-key in Windows Server 2003


Most errors of this type (works against W2K, but not W2K3), that I have
heard of, had to do with the 2003 server trying to switch to TCP because the
packet has become too big.  It seems that either they are putting more in
the PAC, or W2K3 tries to switch earlier than W2K did.

K.C.


-----Original Message-----
From: kerberos-bounces at mit.edu [mailto:kerberos-bounces at mit.edu] On Behalf
Of Ben Cox
Sent: Thursday, June 26, 2003 4:49 PM
To: kerberos at mit.edu
Subject: string-to-key in Windows Server 2003

We are attempting to authenticate against a Windows Server 2003 Active
Directory Server and getting a preauthentication failure (preauth is
enc-timestamp).  We're using a key that we generated from the password
and stored into a keytab; this works against a Win2K AD server but not
against a 2003 server.

Did the string-to-key algorithm change in Win 2003?  (Or does it use a
different mechanism for generating the salt?)

Any info (or pointers to info) on this would be appreciated.

-- Ben


________________________________________________
Kerberos mailing list           Kerberos at mit.edu
https://mailman.mit.edu/mailman/listinfo/kerberos


________________________________________________
Kerberos mailing list           Kerberos at mit.edu
https://mailman.mit.edu/mailman/listinfo/kerberos