[e2e] Two questions on simulation of AQM

Mark Claypool claypool at cs.wpi.edu
Wed Apr 11 06:23:49 PDT 2001

Zhang Miao writes:
 > (1) Where should the action(Marking/Dropping) be taken?
 > (2) What policy should be used in the reverse direction?

In regards to the first question, the following paper is a nice
look at the mark-front approach applied to ECN:

  Chunlei Liu, Raj Jain, "Improving Explicit Congestion Notification
  with the Mark-Front Strategy", Computer Networks, Vol 35, no 2-3, pp
  185-201, January 2001. 
  URL: http://www.cis.ohio-state.edu/~jain/papers/ecnfront.htm

In regards to the second question, we have looked at the effects of
ack loss on TCP performance (typical dumbell topology, high congestion
on back-link).  Only for _very_ high loss rates (about 40%) or very
large round-trip times or small window sizes does ack loss seem to
adversely affect performance.  


 Mark Claypool        CS Assistant Professor        claypool at cs.wpi.edu
   Worcester Polytechnic Institute   http://www.cs.wpi.edu/~claypool

More information about the end2end-interest mailing list