40Kbyte might not be large, but it's enough to fit ~26 1500 byte segments into the pipe, which according to the Westwood documents/publications I have read should be plenty. Dell'Aera et. al. states in their paper that when the pipe size is larger than a few MSS Westwood+ shows improvement over Reno up to 53%. It looks like Linux is using Vegas so I'm not sure how much better Vegas does over Reno which might deplete the difference I am seeing. Of course they also show results much less spectacular, where Westwood+ actually underperformed Reno by -9% at times. In this case the pipe size only accommodated between 0.5-5 MSS. So I believe my BDP should be sufficient, but either way I will increase it and see what happens ... Tx, Adam > -----Original Message----- > From: Filipe Abrantes [mailto:fla@xxxxxxxxxxxxx] > Sent: Tuesday, August 30, 2005 10:16 AM > To: Lewis Adam-CAL022 > Cc: linux-net@xxxxxxxxxxxxxxx > Subject: Re: Westwood performance? > > > > > Lewis Adam-CAL022 wrote: > > Hi all, > > > > I am evaluating the Westwood protocol in the lab using > NISTnet as an > > emulator. So far I am seeing little to zero improvement of TCP > > performance using Westwood as opposed to the Vegas > implementation in > > the kernel. I have two NISTnet boxes to simulate packet loss and > > delay in both the upstream and downstream path. NISTnet is > programmed > > to induce 500ms delay, drop 5% of packets and emulate 1Mbps > bandwidth. > > I'm using iperf and sending 3.2M of data. I am using the 2.6.12 > > kernel. These results are very disappointing as publications I've > > seen show much better improvement with a high BDP (I assume > > bandwidth=1Mbps * delay=500ms qualifies). I am seeing best case 5% > > improvement over Reno/Vegas. > > I am not that familiar with Westwood so I can't tell you what > type of improvement you should expect, but the BDP you > mention is not that high (1Mb*0.5=500kbit=40Kbyte window). > Have you tried larger BDP's with the same result? > > Best Regards > > Filipe > > > > Am I using a valid version of Westwood in this kernel? Is there > > anyway to verify that it is running, aside from cat-ing the /proc > > files? Anything in /var/log/messages or elsewhere? Or is > this really > > the best one can hope for with Westwood (e.g. 5% improvement)? > > > > > > Tx, > > Adam > > - > > : send the line "unsubscribe > linux-net" > > in the body of a message to majordomo@xxxxxxxxxxxxxxx More > majordomo > > info at http://vger.kernel.org/majordomo-info.html > > > > -- > Filipe Lameiro Abrantes > INESC Porto > Campus da FEUP > Rua Dr. Roberto Frias, 378 > 4200-465 Porto > Portugal > > Phone: +351 22 209 4266 > E-mail: fla@xxxxxxxxxxxxx > - : send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
