Testing MapReduce Programs: A Mapping Study

Abstract

The MapReduce (MR) programming model is widely used in the development of parallel applications involving large data sets. Many research papers focus on quality attributes associated to the environments that execute MR programs. However, the test of such programs can be considered a challenger and an open research subject. Considering this fact, this paper presents results from a mapping study about testing of MR programs. The mapping searched by papers related to MR program test, which were analyzed according to the following dimensions: purpose, tools and environments, testing techniques, and criteria and types of fault. The achieved outcomes serve as motivation to stimulate and direct future research in this field.