I have been involved in running an small Internet Service Provider "just for fun". 

One of the things we do is run a usenet server (hosts discussion groups,
as archived at http://google.groups.com/). news servers typically handle 
*lots* of traffic, but its very hard to get useful
statistics out of it, mainly (a) because of the traffic volume
and (b) because news administrators a busy people and never
got round to writing a statistics package for it.

- It's very hard to get decent statistics out of a usenet
server, due to the sheer volume of data that goes through
them (some 400 gig a day on binary transit servers, but
only a few gig a day on a text server).

- the diablo usenet transit server can however just send
a "header feed" (i.e. just the article headers, and not the
main body of the message) to a "server".

This is typically used for propagating cache servers
which then only request the main article when a client
wants it.

Problem 1: Write a daemon which accepts an NNTP
header feed from a news feeder and logs interesting
information into an sql database (group, poster, article
size, date, path, etc.)

Problem 2: Optimise said database, so that querying
multi million row record sets isn't a problem.

Problem 3: Do a web based interface which does tables
and graphs to view data from the database so that people
(news administrators) can view volume of news by
         - group
         - news peer
         - time
         - poster

Also, allow them to identify "top peers" (servers which
most articles go through) and shortest / longest paths.

Some of my ISP colleagues will happily be the client and provide 
a test server & data gathering facilities. I suggest the server 
be written in java or perl. The web interface could be jsp
but might also be done in php.

Student would need to understand unix command lines and bit a bit net savvy.

I can actually see a /huge/ real world use for this package
if it works, cos currently we're all shifting this huge amount
of data around blind and the technique of taking a header
feed would mean it's not nntp server specific.

-  I can immediately think of a user base who could test it
and provide feedback, if user feedback is an issue/requirement.

-- 
peter gradwell (postgraduate) http://peter.gradwell.com/
Dept of Computer Science, Univ. Bath, Bath, UK