[Iccrg] benchmarking new tcp congestion control algorithms

[Mark Allman]

> All around I think a standard battery of tests and some goals for the
> tests based on the tmrg-metrics work would be fantastically useful.

Good idea.  Several comments ...

(1) A strong note should be made that such a battery of generic and
    simple tests should not be the *only* tests one runs.  I.e., they
    should be considered a good start, not a complete set.

(2) We need to make sure that schemes are not tailored to the test
    scenarios.  Perhaps someone would actively try to do this.  However,
    my bigger concern is that such a set of generic simulations would
    lock folks into particular mental models of networks that would lead
    them to implicitly design towards the scenarios and not for broader
    situations. 

(3) The test scenarios need not be completely static.  That is, the RTTs
    / cross traffic / packet corruption rate / reordering / whatever
    should probably be varied randomly (with a number of trials of
    course).

My two bits for a Friday.

allman



-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 185 bytes
Desc: not available
Url : http://oakham.cs.ucl.ac.uk/pipermail/iccrg/attachments/20070309/4814fb07/attachment.bin