hi all, I am new at Bufferbloat and my apologies if I bring up
things that have been already discussed.
We have written two research papers that I think are very relevant
to Bufferbloat:
1. Router Buffer Sizing for TCP Traffic and the Role of the Output/Input
Capacity Ratio,
http://www.cc.gatech.edu/fac/Constantinos.Dovrolis/Papers/buffers-ton.pdf
(with Ravi Prasad and Marina Thottan)
2. End-to-end Detection of ISP Traffic Shaping using
Active and Passive Methods
http://www.cc.gatech.edu/fac/Constantinos.Dovrolis/Papers/shaperprobe-oct10.pdf
(with Partha Kanuparthy)
The first paper revisits some old ("you need bandwidth-delay product
worth of buffering") and more recent ("you only need very short
buffers") buffer sizing rules. It brings up the importance of the
"output/input capacity ratio", which is the ratio between the
capacity of the link that we focus on and the max rate of the
arriving traffic in that link.
The second paper describes our tool "ShaperProbe" (available on M-Lab)
and it describes measurements from many thousands of users at
various ISPs. It gives an idea about the extent of traffic shaping
out there, and the shaping rate & max-burst-size parameters that ISPs
typically use.
We will be glad to participate more actively in BufferBloat, if you
think we can help somehow.
--
Constantine
--------------------------------------------------------------
Constantine Dovrolis, Associate Professor
College of Computing, Georgia Institute of Technology
3346 KACB, 404-385-4205, [email protected]
http://www.cc.gatech.edu/~dovrolis/
_______________________________________________
Bloat mailing list
[email protected]
https://lists.bufferbloat.net/listinfo/bloat
