Ars explains bandwidth and latency
Ars has whipped up another excellent PYAITK article; this time they've tackled
the finer points of the always-important relationship between bandwidth and latency.
This article was written in such a manner that the concepts communicated in it will be applicable to understanding a wide range of systems and parts of systems: from the frontside and memory buses of current P4 and Athlon systems to the buses in Apple's new XServe. Throughout the article, then, I've deliberately avoided getting mired down in the details of specific implementations in hopes that the general concepts will stand out clearly for the reader.
In usual Ars style, the article isn't light reading, but the generalized approach makes it a valuable resource for anyone looking for some insight on the subject. You'd be well-advised to make sure you're fully awake before tackling this one.