This paper presents a MapReduce algorithm for computing pairwise document similarity in large document collections.