Re: draft-ietf-dnsext-dnssec-gost

Paul Hoffman <> Thu, 11 February 2010 21:23 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id D37B228C164 for <>; Thu, 11 Feb 2010 13:23:28 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -6.033
X-Spam-Status: No, score=-6.033 tagged_above=-999 required=5 tests=[AWL=0.013, BAYES_00=-2.599, HELO_MISMATCH_COM=0.553, RCVD_IN_DNSWL_MED=-4]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id e2UkiAJ5VtU4 for <>; Thu, 11 Feb 2010 13:23:28 -0800 (PST)
Received: from (Balder-227.Proper.COM []) by (Postfix) with ESMTP id 2A0C328C169 for <>; Thu, 11 Feb 2010 13:23:28 -0800 (PST)
Received: from [] ( []) (authenticated bits=0) by (8.14.2/8.14.2) with ESMTP id o1BLOKBQ044792 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 11 Feb 2010 14:24:20 -0700 (MST) (envelope-from
Mime-Version: 1.0
Message-Id: <p06240804c79a234f3ad8@[]>
In-Reply-To: <>
References: <p06240806c799d87e7406@[]> <> <>
Date: Thu, 11 Feb 2010 13:24:05 -0800
To: Andrew Sullivan <>, Olafur Gudmundsson <>
From: Paul Hoffman <>
Subject: Re: draft-ietf-dnsext-dnssec-gost
Content-Type: text/plain; charset="us-ascii"
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: IETF-Discussion <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 11 Feb 2010 21:23:28 -0000

At 4:04 PM -0500 2/11/10, Andrew Sullivan wrote:
>So the question here is not what algorithms get "first class" status
>in general, but whether we want to have different classes of support
>for DNSSEC, given the current conditions. 

First off, thank you for better stating the question.

There are a plethora of signing algorithms. Note that a signing algorithm consists of a public key algorithm *and* a hash algorithm.

The question here is whether they also have SHOULD-level requirements to process every signing algorithm that is in the IANA registry. Having such a requirement gives attackers a much wider target: in order to spoof a signature, they can pick the weakest of a large collection of algorithms.

For example, there is already a published attack on the GOST hash function that does not exist in SHA-256 and SHA-512. The GOST algorithms have had much less cryptographic review than other algorithms. If that attack becomes practical, an attacker can create signatures using GOST that he/she could not create in RSA/SHA-256 or RSA/SHA-512.

Given this, the answer to the question should be "no, not all algorithms automatically get SHOULD-level requirements". The IETF can, on a case-by-case basis, decide if they want to update the base DNSSEC spec to include a SHOULD-level or MUST-level requirement for a new signature algorithm.

--Paul Hoffman, Director
--VPN Consortium