On Wed, 25 Sep 2002, Mazhar Memon wrote: > I'm sure this has been discussed before, so deal with it. I have a > modem connection and since a lot of the throughput penalty is due to > the enormous transmitting and receiving latency and since the modem > link is pt-to-pt it seems that I can really benefit from jumbo mtu > sizes. I was a little confused when I saw the follow-ups before your original message. This is a good example of needing to know the real problem before fine-tuning an assumed solution. TCP is a windowed protocol, thus increasing the MTU size will not help. In your case increasing the MTU will likely further increase effective latency because of a reduced chance to prioritize packets and higher cost of dropped frames. The only case where increasing the MTU helps is when - the extra bytes in the protocol header impacts effective bandwidth, or - the CPU cost of handling extra packets or headers is high. > It seems that this could jave be done before since its such > a simple idea. While breakthroughs sometimes happen with "why didn't someone think of this before", most of the time the answer is "they did, and rejected the idea because they understood the real problem." -- Donald Becker becker@scyld.com Scyld Computing Corporation http://www.scyld.com 410 Severn Ave. Suite 210 Second Generation Beowulf Clusters Annapolis MD 21403 410-990-9993 - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
