data sets, measurement sites

John Snell (geigudr@cs.washington.edu)
Mon, 3 Aug 1998 10:04:53 -0700 (PDT)

Tom requested that I put up a list of other measurement sites. As the
Sirpa-* is down, it's on remus.

http://remus/detour/komatos.html

Any additional sites will be added as mentioned to me.

Additionally, I finished polishing the data set. What has been removed:

1. Empty records. I'm still tracking down the reason for this.
2. Dupes -- only 15 dupes, which came to something like 0.01% of
the final american data set.
3. IP's that changed during the measurement were standardized
to one IP (two of them).
4. Robots.txt removals were removed, since they had very few
results in the set.
5. A->A paths. I thought I'd disabled this behavior, but
apparently not. The funny thing, is to look at the amount of
packet loss some of these machines got in tracerouting to
themselves.

All in all, the size of the uncompressed dataset is now 89MB, from 94MB.

I recommend that all users of the dataset use this particular version.
The original files will be available in an archive called
"oldAmerica.zip", same directories.

Additionally^2, I've compressed the data files. So now the d/l is only
12 meg.

http://remus/phase1/america.fix.zip
ftp://romulus/pub/phase1/america.fix.zip
...

And One other thing; it looks like we might have a feasible set of sites
to do an Asia/Pacific rim study, along with Europe. So far, I have
servers in:

Australia (mostly)
Japan
Korea
Hong Kong
Korea
New Zealand
Phillipines

And I'm still working through Neal's list, so this may well expand.
Something of a tradeoff; I still haven't done anything interesting with
the EuroSlave data, and most people only care about America. But
something to consider.

_________________________________________________________________________
50 USC Sec. 1520:
(b)(1) The Secretary of Defense may not conduct any test or experiment
involving the use of any chemical or biological agent on civilian
populations unless local civilian officials in the area in which the test
or experiment is to be conducted are notified in advance of such test or
experiment, and such test or experiment may then be conducted only after
the expiration of the thirty-day period beginning on the date of such
notification.