Hadoop at home

Abstract

The potential benefits of data-intensive scalable computing (DISC) in CS education are considered in the context of a small college with an active student-operated Beowulf cluster initiative. The map-reduce computational model, of great importance in industry, is reviewed, and the Hadoop implementation of that model is connected to specific courses throughout the undergraduate CS curriculum. Concerns when running a local Hadoop-capable cluster at a small college are identified.