Bugzilla – Bug 392
Example star.cc confuses TCP
Last modified: 2008-12-06 03:13:35 EST
Change the ns3::OnOffApplication::DataRate on line 54 of example/star.cc to 15kb/s as in, Config::SetDefault ("ns3::OnOffApplication::DataRate", StringValue("15kb/s")); TCP gets confused, and the simulation doesn't stop. A tcpdump shows (at the supposed end end of the simulation) 9.843352 IP 10.1.1.2.49153 > 10.1.1.1.50000: . 16441:16578(137) win 65535 9.916419 IP 10.1.1.2.49153 > 10.1.1.1.50000: . 16578:16715(137) win 65535 9.916419 IP 10.1.1.1.50000 > 10.1.1.2.49153: . ack 16715 win 65535 9.989486 IP 10.1.1.2.49153 > 10.1.1.1.50000: . 16715:16852(137) win 65535 10.000000 IP 10.1.1.1.50000 > 10.1.1.2.49153: F 1:1(0) win 65535 10.004134 IP 10.1.1.2.49153 > 10.1.1.1.50000: . ack 2 win 65535 10.144326 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 10.144326 IP 10.1.1.1.50000 > 10.1.1.2.49153: . ack 16853 win 65535 10.189486 IP 10.1.1.1.50000 > 10.1.1.2.49153: . ack 16853 win 65535 10.454007 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 11.073370 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 12.312094 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 14.789542 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 19.744439 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 29.654233 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 39.564027 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 49.473821 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 59.383616 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 69.293410 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 79.203204 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 89.112998 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 99.022792 IP 10.1.1.2.49153 > 10.1.1.1.50000: F 16715:16852(137) win 65535 and this goes on "forever" ...
Created attachment 322 [details] Proposed fix This fixes the issue, but I require some feedback from George as to the "correctness" of the fix. I hope to have an answer this afternoon.
George agrees that this looks good. I am doing some testing to make sure that this doesn't break things...