Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 11 Mar 2005 08:27:36 -0500
From:      Bart Silverstrim <bsilver@chrononomicon.com>
To:        freebsd-questions@freebsd.org
Subject:   Re: Clock slew vulnerability in FreeBSD?
Message-ID:  <56f756c499c68c62c6706fef0e896cb2@chrononomicon.com>
In-Reply-To: <1735368246.20050311044408@wanadoo.fr>
References:  <751280160.20050311034539@wanadoo.fr> <20050311025906.GD72527@hub.freebsd.org> <1735368246.20050311044408@wanadoo.fr>

next in thread | previous in thread | raw e-mail | index | archive | help

On Mar 10, 2005, at 10:44 PM, Anthony Atkielski wrote:

> Kris Kennaway writes:
>
>> Isn't this a non-problem if you use ntpd?
>
> Unfortunately, no, because the TCP stacks on most systems don't use the
> disciplined clock provided by NTP for the timestamps.  Instead they use
> a clock based directly on the RTC, which reveals a characteristic skew
> that is unique to each machine.
>
> If the stacks used the NTP-disciplined actual time of day, plus perhaps
> a randomizing factor to avoid revealing patterns, this technique would
> become useless.

Wouldn't the skew resolution necessary for this tracking technique 
become useless with temperature variations, humidity, etc. that can 
affect most systems over the course of the day/week/year?



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56f756c499c68c62c6706fef0e896cb2>