http://remus/phase1/
ftp://romulus/phase1/
Note the use of the term "phase1". The original measurement period ended
just before the retreat, due to server bugs. The next period has not yet
begun. This period lasted from May24-June28. (approx)
You'll find a dat file and an lst file.
The dat file is in a specialized format, but easily understandable -- it's
a generalized form of traceroute to allow for use of both NT and Unix
traceroute variants. The dat file is also sorted according to symmetric
path. All times are in microseconds, hence the use of integers.
The line:
RoutedIP: 12@206.205.230.81:164244,0;187278,0;143587,0
Means,
at TTL = 12, 206.205.230.81 responded 3 times, with
164244,187278,143587 mics as times, and all those times had a non-error
status code (the 0).
There are other status codes; they generally match those you've already
seen in traceroute (!H, !N, !A, !, *, etc).
The lst file is a list of paths traced; the list file matches up to the
dat file in the following manner:
.lst: path A -> B
.lst: path C -> D
.dat: path A -> B
.dat: path B -> A
.dat: path C -> D
.dat: path D -> C
Probewait is assumed to be the default specified; it says 3000
(milliseconds), Paxson's thesis claims that should be 5000, despite what
the docs say. My stopwatch agrees.
Tracewait is the duration that the trace CGI was allowed to complete its
task; that's also in ms. Start/End time are seconds since 1900, according
to RFC 868 (ftp://ftp.isi.edu/in-notes/rfc868.txt). The timeServerIP is
where that time was obtained from (there are several used, round-robin).
Anyone desiring Java code that loads all this up into a hashtable/vector
setup (keyed access) should send specific mail; I'll add you an account
into the source server (NT/VSS), and run you through the basics of using
the code.
_____________________________________________________________________________
"The human mind is a 400,000-year-old legacy application...and you expected
to find structured programming?" -- Randall Davis, 1996 AAAI Pres. Address