Cluster computing!

Computing in Science and Engineering (published as Computational Science and Engineering prior to 1999) is a joint publication of IEEE Computer Society and American Institute of Physics. In my opinion, it is the computing equivalent of American Journal of Physics (for physics) and American Mathematical Monthly (for mathematics): a journal that addresses issues related to the culture and pedagogy of scientific computing. A recent issue - Volume 7, Issue 2, Year 2005 - of this journal carries several interesting articles on cluster computing: the articles are extremely accessible. Questions like "What is a cluster?", "Is Linux the operating system of choice for cluster computing?" are answered in the guest editorial. Apparently, the articles for this special issue were solicited with the aim of instructing (a) how to get started with cluster computing (especially a novice, who had never done it before), (b) how to get the most out of cluster computing, and (c) how to develop serious applications that scale well. There are four articles: one on Beowulf clusters, one on plug-and-play cluster computing, one on the use of Java for cluster computing, and finally, one on resource-aware scientific computation on a heterogeneous cluster. While you are at it, you may also want to take a look at their "Perspectives in Computational Science" column on high-performance computing.