[e2e] Question on assumption of congestion control algorithm

Sally Floyd floyd at aciri.org
Mon May 7 08:17:05 PDT 2001


>tweaking TCP algorithms
>to recognise and cope with reordering (e.g. by selectively raising
>the dupack threshold) is an area I haven't seen much written about
>yet.

Yep.  Hopefully there will be more shortly.
There is some preliminary work in this direction in the DSACK standard:

An Extension to the Selective Acknowledgement (SACK) Option for TCP. 
Floyd, S., Mahdavi, J., Mathis, M., and Podolsky, M.. 
RFC 2883, Proposed Standard, July 2000. 
ftp://ftp.isi.edu/in-notes/rfc2883.txt

And some discussion of possible paths for making TCP robust to 
reordering in the following:

A Report on Some Recent Developments in TCP Congestion Control,
Floyd, S. IEEE Communications Magazine, April 2001. 
http://www.aciri.org/floyd/papers/report_Jan01.ps

- Sally




More information about the end2end-interest mailing list