On 8/27/06, David Miller <davem@xxxxxxxxxxxxx> wrote:
> > I would love 3, 4, 5 to go into 2.6.18 as these resolve long standing CCID3 > issues that have been in the DCCP tree since inception and have caught a > number of people. Ok, I'll toss 1-5 into 2.6.18
Thanks for that. Are 6 and 7 going into 2.6.19 or do you want Arnaldo to have a bit more of a look? 6 in particular is trivial.
One thing I don't understand is this description from patch 5: -------------------- This gives a theoretical speed of 71.9 Kbits/s. I measured across three runs with this patch set and got 70.1 Kbits/s. Without this patchset the average was 232 Kbits/s which means Linux can't be used for CCID3 research properly. -------------------- Decreasing the transfer rate is desirable? I read this as saying this "fix" drops the transfer rate down from 232Kb/sec to 70.1Kb/sec. What's going on here?
DCCP CCID3 (RFC 4342) uses TFRC (RFC 3448) to calculate the desired rate to send at based on feedback from the receiver. The reason for this is that TFRC is not ACK/Window based to control rate and TFRC calculates a rate so that the flow is "fair" when competing with TCP. TFRC is designed to be smoother than TCP at dealing with loss - more sine wave than saw tooth. The calculation is based on the work Padhye et al did in this paper - http://citeseer.ist.psu.edu/padhye98modeling.html As it turns out this is based on TCP Reno at that time and modern TCP variants are more efficient when dealing with loss as can be verified through iperf but we should implement what the RFC says. Basically the implementation in the DCCP code was buggy and was transmitting too fast so I have made it conform to the RFC much closer. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line "unsubscribe dccp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
