Dear all, I wanted to suggest something, hoping that it might become a work item. Actually I had hoped to present it in Stockholm... I'm not sure if one of us (this is my Ph.D. student Dragana Damjanovic's work) can make it to Hiroshima, but probably yes. Our proposal is "MulTFRC", which is... well, I guess the name says it all. Anyway it seems to us to do a better job than MulTCP or any other such protocol, and it works quite well with a wide range of emulated flows from e.g. 0.1 to 100. It is also more reactive and smoother than multiple individual TFRC flows would be. We just had a paper on it accepted, which should appear in the next issue of CCR: http://www.welzl.at/research/publications/ccr-multfrc.pdf and we have simulation code: http://dps.uibk.ac.at/~dragana/mulTFRC.html as well as real life code which we will make available soon. What do you think? Cheers, Michael > DCCP folks, > > We have been planning that DCCP will not meet this time in Stockholm. > Please let us know if this is not ok with you. The currently active > drafts will soon be completed, so we'd like to encourage people to > propose new candidate DCCP work items, to be discussed in Hiroshima. > > Thanks! > > - Pasi & Tom > >
