On a Hadoop-based Analytics Service System

In this study, we discuss the development of an analysis service that uses Hadoop for improving the performance of a database management system (DBMS)-based analysis service system that processes big data. DBMS-based systems are not suitable for processing big data and providing service because of their disadvantage in consuming more time for processing and analyzing. We introduced a distributed parallel platform, Hadoop ecosystem, for improving the performance of the system by minimizing the processing time in analyzing big data. In addition, we also carried out a method of optimizing an existing analysis module to minimize the processing time. SPARK and Hive were implemented in the Hadoop platform to optimize the distributed parallel infrastructure. The developed analysis service system showed 127.87 times higher performance than that of other existing systems.