Everyone has heard about the Slashdot Effect, but I have yet to find a post anywhere on the Web showing a concrete, quantified example of what exactly happens to a web server when the Slashdot hordes attack it. Here, then, is a snip of the EWS web server logs from UIUC (courtesy of my friend Pete):

(The following statistics have been edited to show only byte totals.)
Slashdot of EWS webserver due to 6/03/2000 7:14 PM headliner
(http://slashdot.org/articles/00/06/03/2256247.shtml).

This list only includes Internet traffic.
The ranking is based on traffic, and is relative to all other UIUC machines for that day.
Date Rank             Bytes %Src %Snk
---- ---- ----------------- ---- ----
5/31   95       429,304,111  91%   8%
6/01   84       436,229,395  92%   7%
6/02   99       370,694,862  91%   8%
6/03   20     3,066,780,013  96%   3%
6/04   14     4,424,011,397  96%   3%
6/05   28     2,351,673,597  95%   4%
6/06   47       895,033,112  92%   7%
6/07   63       696,557,513  93%   6%
6/08   72       510,076,735  92%   7%
6/09   75       484,958,942  92%   7%

This indicates a ten-fold increase in traffic (on the day after the article posting) from standard College of Engineering web traffic. I'm happy to say that EWS didn't bat an eyelash at it.

Update: Let me also note that the webpage in question contained a small amount of text and a few pictures -- nothing that would normally cause a lot of bandwidth usage. That gives you some idea of the sheer volume of hits involved.