15. On Performance of Commodity Single Board Computer-Based Clusters: A Big Data Perspective

In recent times, the commodity Single Board Computers (SBCs) have now become sufficiently powerful that they can run standard operating systems and mainstream workloads. In this chapter, we investigate the design and implementation of Single Board Computer (SBC)-based Hadoop clusters. We provide a compact design layout and build two clusters each using 20 nodes. We extensively test the performance of these clusters using popular performance benchmarks for task execution time, memory/storage utilization, network throughput, and energy consumption. We investigate the cost of operating SBC-based clusters by correlating energy utilization for the execution time of various benchmarks using workloads of different sizes. Although the low-cost benefit of a cluster built with ARM-based SBCs is desirable, these clusters yield low comparable performance and energy efficiency due to limited onboard capabilities. It is, however, possible to tweak Hadoop configuration parameters to ARM-based SBC specifications to efficiently utilize available resources. Finally, a discussion on the design implications of these clusters as a testbed for inexpensive and green cloud computing research is presented.