Hadoop is all about processing large amounts of data (so-called Big Data) across a cluster of nodes. In this case, the nodes are Raspberry Pis and Nigel Pond has written a PDF tutorial on how to do it. Read it here.