Linux routing project

The information presented here is a running documentation of an ongoing project and is currently intended for internal use only. Please do not create links pointing to this page. If you are interested in referencing any of the information presented here, please contact us at lin@cs.unh.edu.

 

Phase 1       Experiment on Static Routing Performance

                 When packet size is optimal, the max throughput can be about 1% less than ideal theoretical throughput 100Mb/sec. See experiment

                  Question 1: Why  the RTT is much less than it was when packet size is 16KB or 64KB when packet size is 32KByte? See Here.

                  Question 2: Why the RTT is NOT that stable when packet size is 64KByte and this fluctuation seems not because of TCP retransmission
                                      See here

Phase  2       Experiment packet dropper on Kernel 2.4.10, Reading Kernel Code and Documents on Advanced Routing Topics

                    Questions: Regular Dropping gets a much better performance than Random Dropping. Why??
Random Dropping gets a much better performance than NS(network simulator) result. Why??
When doing regular dropping, if we set dropping rate to 1 out of every 6 (or less), The throughput varies significantly each time. why???

Phase 3        Coding & debugging & testing

Useful Resources