Re: iSCSI: DH-CHAP (correction)

["Ofer Biran" <BIRAN@il.ibm.com>]

A correction to my previous post:

"...should not cause the WG to change for something technically deficient."

The "technically deficient" phrase is an error - what I wanted to say is -
technically inferior due the deficiency mentioned.

  Regards,
    Ofer


Ofer Biran
Storage and Systems Technology
IBM Research Lab in Haifa
biran@il.ibm.com  972-4-8296253


---------------------- Forwarded by Ofer Biran/Haifa/IBM on 13/04/2002
23:31 ---------------------------

Ofer Biran/Haifa/IBM@IBMIL@ece.cmu.edu on 13/04/2002 17:44:56

Please respond to Ofer Biran/Haifa/IBM@IBMIL

Sent by:    owner-ips@ece.cmu.edu


To:    David Jablon <dpj@theworld.com>
cc:    David Black <Black_David@emc.com>, <ips@ece.cmu.edu>,
       <ElizabethRodriguez@ieee.org>, <Elizabeth.G.Rodriguez@123mail.net>
Subject:    Re: iSCSI: DH-CHAP




David,

Being that 'participant' I'd like to clarify that my comments (that
were not that private as the security team was copied) were on an
earlier rough draft David Black posted to the security team just few days
before posting to the general IPS list, and apart from that I'm not aware
of '[closed] design process' on the security team for DH-CHAP.

My main comment was about active impersonation + off line dictionary
attack and a misleading text (in my view) that ignored this attack. The
final version now clearly describes it both in the overview and section 6.

I also commented that getting a password can cause much more damage
than connection hijack after login phase, and this is also mentioned in
section 6.5.

So one has to admit that the draft states fairly and clearly the main
DH-CHAP deficiency - vulnerability to active dictionary attack.

Now - the WG should decide whether the 'IP issue' of SRP is a good enough
reason to replace it with another mandatory method, introducing this
deficiency. SRP was originally chosen over CHAP due to the risk of an
attacker obtaining the password. DH-CHAP only makes that attack 'networkly'
more difficult, but still possible. As I understand it, the IP situation
of SRP (free license of the actual patent, 'reasonable and
non-discriminatory' IETF statements for the patents that were brought up as
'might be related'), according to the IETF policy, should not cause the WG
to change for something technically deficient.

I currently vote for putting DH-CHAP as another MAY method (it does provide
valuable resilience over CHAP in certain environments, and the draft seems
in a pretty good shape), unless somebody convince me that I misunderstood
the
SRP IP situation and/or the IETF policy.


   Regards,
     Ofer


Ofer Biran
Storage and Systems Technology
IBM Research Lab in Haifa
biran@il.ibm.com  972-4-8296253