I sent in the revised version of draft-ietf-dccp-tfrc-voip-05.txt this morning, in response to feedback from Working Group Last Call. A local copy is available at:
http://www.icir.org/floyd/papers/draft-ietf-dccp-tfrc-voip-05.txt http://www.icir.org/floyd/papers/draft-ietf-dccp-tfrc-voip-05.ps Figures for the graphs are available at: http://www.icir.org/tfrc/voipsims-05/voipimages-05.ps http://www.icir.org/tfrc/voipsims-05/voipimages-05.pdf The list of changes is appended below. The most interesting additions are the last two tables in Section 7.3. "These two sets of simulations illustrate the dangers of a widespread deployment of TFRC-SP in an environment where small packets are less likely to be dropped than large ones." - Sally http://www.icir.org/floyd/ Changes from draft-ietf-dccp-tfrc-voip-04.txt: * Added tables showing the response function for TCP, TFRC, and TFRC-SP, for a range of packet sizes, for both packet and byte drop rates. In response to feedback from Magnus Westerlund. * Along with response function, added that TCP's sending rate varies linearly with packet size. From a suggestion by Magnus Westerlund. * Added a sentence saying that TCP has a range of sender algorithms for setting the RTO. * Deleted the sentence equating TFRC-SP with TFRC-PS referred to in RFC 3448. From a suggestion by Colin Perkins. * Added that wireless links sometimes are less likely to drop small packets. Reported from Pete Sholander. * Added simulations to the end of Section 7.3 comparing the effects of TFRC and of TFRC-SP, for an environment with a Drop-Tail queue in bytes, showing the possible negative consequences of TFRC-SP. In response to email from Magnus Westerlund. * Added an explanation for the Adaptive RED simulation with packet drop rates greater than 50%. In response to email from Magnus Westerlund. * Added a Conclusions section, with a sentence that a separate document will be used to specify an experimental CCID based on TFRC-PS. In response to feedback during Working Group Last Call. * Added a paragraph about "Initializing the Loss History after the First Loss Event" in TFRC-SP.
