I downloaded one to play with it, and hacked their scripts to print out
TCP packets in tcpdump format, if you want to check it out:
[DETOUR]% pwd
/data1/nlanr/data/19981130
[DETOUR]% /data1/nlanr/bin/crl2asc.pl crl-TXS-912460627 | less
I think i might try to use their stuff for the next rev of the tcp
modeling stuff, since it's 1) real, 2) fresh, 3) diverse. The drawback of
this approach is that i may have to filter down the data to flows where we
see the sender side (so we can determine the loss rate), and the sender is
probably TCP-limited (as opposed to disk or CPU limited).
neal
---------- Forwarded message ----------
Date: Tue, 1 Dec 1998 08:25:48 -0800
From: Hans-Werner Braun <hwb@nlanr.net>
To: end2end-interest@ISI.EDU
Subject: data
In case people are interested, there is a lot more network data in
these days than there was, and also preparsed stuff in
http://moat.nlanr.net/Datacube
Obviously the documentation is lagging behind....
More general info in
and
http://moat.nlanr.net/Presentations/NAI