Re: TCP and UDP window size?

To: iperf-users --at-- dast.nlanr.net, lee.langley --at-- ntlworld.com
Subject: Re: TCP and UDP window size?
From: "Paul Hyder" <Paul.Hyder --at-- noaa.gov>
Date: Wed, 23 Mar 2005 14:55:26 -0700
Content-transfer-encoding: 7bit
Content-type: text/plain; charset=us-ascii; format=flowed
In-reply-to: <001101c52f95$afd9db50$66fefe0a --at-- synetrixhl.local>
References: <001101c52f95$afd9db50$66fefe0a --at-- synetrixhl.local>
Reply-to: iperf-users --at-- dast.nlanr.net
Sender: owner-iperf-users --at-- dast.nlanr.net
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3) Gecko/20040913

The TCP window needs to be adjusted based on any connection delay due to distance.
If you are in the same building 128K is fine.  At a round trip time of say 100ms
a much larger tcp window is necessary for full speed transfers (see the formula).

Your 125K should be ok for RTT times less than ~10ms at 100Mbps.
	Paul Hyder

Lee Langley wrote:
> Hello every one,
> 
> I have a question, I am reading through the user documentation for IPERF and wanted some clarification.
> 
> I am typically using Iperf to measure available true bandwidth through switching products at 100mbps, now I want to understand what the best packet window size would be for such a test, so far I have just assume a window size of 128K....
> I see there is an equation in the documentation for calculating this but do you really need to be this specific, am I getting the correct results sticking with a 128k packet window ?
> 
> Thank you for your feedback.
> 
> Lee

References:
- TCP and UDP window size?
  - From: "Lee Langley"

Prev by Date: Iperf 2.0.1 - multicast problem
Next by Date: DAST: Iperf - robertdixon --at-- avaya.com
Previous by thread: TCP and UDP window size?
Next by thread: Iperf 2.0.1 - multicast problem

Other Mailing lists | Author Index | Date Index | Subject Index | Thread Index