[83788] in North American Network Operators' Group
Re: Iperf or Iperf like test points?
daemon@ATHENA.MIT.EDU (Mark Boolootian)
Thu Aug 25 09:44:02 2005
Date: Thu, 25 Aug 2005 06:43:05 -0700
From: Mark Boolootian <booloo@ucsc.edu>
To: Jared Mauch <jared@puck.nether.net>
Cc: "Walter C. Ames" <walta@mdconnect.com>, nanog@nanog.org
Reply-To: booloo@ucsc.edu
In-Reply-To: <20050825131309.GI68966@puck.nether.net>
Errors-To: owner-nanog@merit.edu
There's one more to add to the list, and it is typically the most
common problem for paths over the wide area: window size must be
adjusted corresponding to the bandwidth*delay product. e.g. if you have
a 20 msec RTT and a 32KByte window, you won't be able to do any better
than ~13 Mbit/sec.
