Squid (named so because all other good names were apparently taken - no, honest!) is a HTTP/Gopher/FTP proxy caching server, mainly intended to be used in UNIXes. Its primary goal of life, as with other proxies of this type, is to fetch stuff off the network and keep a copy in local network, in case someone there wants it really quickly. At the moment, it is likely to be one of the most widely used proxy servers. It is distributed under GPL.
It was originally developed by National Laboratory for Applied Networking Research (NLANR), now also by volunteers. NLANR's Duane Wessels is the current project maintainer.
I have used Squid for some time on my own machine to speed up ISDN web browsing, and it helped somewhat even when it's actually much bigger than what's needed for this purpose. Now that I have broadband connection it's probably redundant, but I have forgotten to uninstall it. =) In Debian the setup is very simple, just about as tricky/trickless as Apache.
Squid is very configurable. It supports many kinds of access control, ICP/HTCP interfacing with other proxies, many sorts of messing with the data (including header mangling a la header stripping and User-Agent stripping)... It keeps things cached in directory, metadata in memory, and also can keep very often needed things in RAM.
Home page: http://www.squid-cache.org/
Sources: SQUID FAQs