[e2e] Re: crippled Internet]

[stanislav shalunov]

Wenyu Jiang <wenyu at cs.columbia.edu> writes:

>   Because I missed the initial part of this discussion, can someone
> please point to me the exact places (urls, etc.) where delays on
> Internet2 is very high (e.g., exceeds 500ms)? I browsed through the
> Surveyor [...]

Be careful when you interpret the results.  The machines that serve as
measurement points aren't on the backbone and the vast majority of
problems is with campus connectivity (much less often GigaPoP, and
exceedingly rare with the backbone per se).

(There exist measurement stations on the backbone but their precision
is not the same as that of the rest and data isn't available through
the normal Surveyor interface.)

If you see any noticeable jitter or loss (say, above few ms and 0.001%
unidirectionally, respectively) you can assume that there's trouble
somewhere and can start to work it out by contacting your local NOC.
More often than not it's something stupid like 100BaseT duplex
mismatch on a campus LAN.

Also see the Internet2 End-to-End Performance initiative at
http://www.internet2.edu/e2eperf/

-- 
Stanislav Shalunov		http://www.internet2.edu/~shalunov/

Beware of Programmers who carry screwdrivers.    -- Leonard Brandwein