Performance Evaluation of Mini Single Board Computer in Hadoop Big Data Cluster

Abstract

Abstract Hadoop is a Java-based and open source software framework that serves to process big size data in a distributed manner. Hadoop uses a framework for applications and programming called MapReduce. Along with the development of Hadoop, big data research using Hadoop is running using a single board computer (SBC). SBC is a mini computer which has a specification similar to a conventional computer but has lower specifications and power usage. However, the use of storage used on a single board computer has constraints, namely low bandwidth and limited local storage, therefore NAS is used because it has large storage and bandwidth speeds is upto1 Gbps. It uses as storage that is carefully used on Hadoop. Based on testing by running Hadoop in a single node and multi nodes using Micro SD storage and NAS does have a significant difference. When running Hadoop on a single node the difference in data access speed generated by Micro SD and NAS has a difference of 2 seconds. On multi nodes (clusters) the ratio is 4 seconds. The difference in NAS bandwidth that is 2 times faster than Micro SD show a significant difference if it is run on Hadoop.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint