Re: [openpgp] Dealing with clock skew

Benjamin Kaduk <> Mon, 18 November 2019 09:50 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 1CB9E1208A9 for <>; Mon, 18 Nov 2019 01:50:14 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -4.2
X-Spam-Status: No, score=-4.2 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, SPF_HELO_NONE=0.001, SPF_PASS=-0.001] autolearn=unavailable autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id A6JSWObUphki for <>; Mon, 18 Nov 2019 01:50:13 -0800 (PST)
Received: from ( []) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 008091208FA for <>; Mon, 18 Nov 2019 01:50:12 -0800 (PST)
Received: from ([]) (authenticated bits=56) (User authenticated as kaduk@ATHENA.MIT.EDU) by (8.14.7/8.12.4) with ESMTP id xAI9o1Zg009631 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 18 Nov 2019 04:50:04 -0500
Date: Mon, 18 Nov 2019 01:50:01 -0800
From: Benjamin Kaduk <>
To: Jon Callas <>
Cc: Claudio Luck <>,, Jon Callas <>
Message-ID: <>
References: <> <> <>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <>
User-Agent: Mutt/1.12.1 (2019-06-15)
Archived-At: <>
Subject: Re: [openpgp] Dealing with clock skew
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: "Ongoing discussion of OpenPGP issues." <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Mon, 18 Nov 2019 09:50:14 -0000

On Sat, Nov 16, 2019 at 03:06:13PM -0800, Jon Callas wrote:
> In the general case, you can't consider a time measurement to be a scalar, it has to be at the very least a complex number of the form [time, skew]. As Derek noted, Kerberos used a skew of five minutes. While Neal Walfield noted in his original post that he's seen skew of 20min, I concur that that seems a bit long. My naive home set-up commonly has alarms across devices being ±2s or less, but that's because they're all getting time from some combination of NTP and cellular network time, which is ultimately GPS time (and of course, skew). I think five minutes is likely reasonable, but *some* skew is unavoidable. Moreover, anyone who's on satellite networks is seeing latency of over a second and once you throw in normal exponential backoff, five minutes seems about as short as is reasonable.

I believe that if Kerberos was starting over now, the 5 minutes would be
seen as excessively long, FWIW.