[IPsec] Comments to the draft-katagi-ipsecme-clefia-01.txt

Tero Kivinen <kivinen@iki.fi> Sun, 04 November 2012 02:13 UTC

MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Message-ID: <20629.53061.77451.107237@fireball.kivinen.iki.fi>
Date: Sun, 04 Nov 2012 04:13:25 +0200
From: Tero Kivinen <kivinen@iki.fi>
To: Masanobu.Katagi@jp.sony.com, ipsec@ietf.org
Subject: [IPsec] Comments to the draft-katagi-ipsecme-clefia-01.txt
Precedence: list

This document defines clefia ciphers for IKEv1 and IKEv2. I see no
point of adding anything to the already obsoleted IKEv1 protocol, so I
think it is better to remove all references to IKEv1 from this
document (i.e section 3) and IKEv1 parts from section 1 and section 6.

We need to be consistent here and reject all additions to IKEv1.

Also some of the IKEv1 registries (IPSEC ESP Transform Identifiers)
are "Standards Track RFC" required registries, meaning this document
would have to be standard track document. The document is currently
listed as being informational.

In the section 2.4 this document adds CLEFIA-CMAC-PRF-128, in similar
way than AES-CMAC-PRF-128 was added. I thought we had already
discussion in the ipsec-list about whether any CMAC-PRF for IKEv2 KDF
use is safe, but I cannot find it now. I think some cryptographers
were saying that CMAC is not good enough for using as KDF (which is
where IKEv2 PRF is used for) as its output is not random enough or
something.

We should really look in to this issue, and see whether we need to
deprecate the other CMAC PRFs too.
-- 
kivinen@iki.fi

[IPsec] Comments to the draft-katagi-ipsecme-clef… Tero Kivinen