[21445] in bugtraq

home help back first fref pref prev next nref lref last post

Re: Small TCP packets == very large overhead == DoS?

daemon@ATHENA.MIT.EDU (Brett Lymn)
Tue Jul 10 11:08:52 2001

From: blymn@baesystems.com.au (Brett Lymn)
Message-Id: <200107101053.UAA06414@mallee.awadi>
To: avalon@coombs.anu.edu.au (Darren Reed)
Date: Tue, 10 Jul 2001 20:23:03 +0930 (CST)
Cc: c3rb3r@hotmail.com (gregory duchemin), bugtraq@securityfocus.com
In-Reply-To: <200107092228.IAA26460@caligula.anu.edu.au> from "Darren Reed" at Jul 10, 2001 08:28:06 AM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

According to Darren Reed:
>
>> stupid window size of one byte may generate a very high overhead.
>
>Silly window sizes aren't so bad.
>

Oh no, they can be very bad and can choke performance measurably.
They are very subtle too because you will not see them if you are
"close" to the server.  Once your RTT increases then you are more
likely to see this problem, especially if you have a fat pipe (eg
cable, adsl) because you can quickly fill the window and then have to
wait for the acks to trickle back.

-- 
===============================================================================
Brett Lymn, Computer Systems Administrator, BAE SYSTEMS
===============================================================================



home help back first fref pref prev next nref lref last post