BIG DATA BENCHMARK PADA HADOOP 2, SPARK, DAN PRESTO MENGGUNAKAN METODE PERBANDINGAN WAKTU RESPON QUERY; BIG DATA BENCHMARK ON HADOOP 2, SPARK, AND PRESTO BASED COMPARISSON ON QUERY RESPONSE TIME

BASKORO, DANIEL OSCAR, Lukman Heryawan

2015 | Skripsi | FMIPA

Along with the improvement of the information in digital form (digital data), the need for analysis of big data becomes a priority for the society, especially the private sector and government. These needs to be a priority for the Big Data analysis can provide information that can facilitate policy-makers in determining a policy. High demand in the analysis of Big Data, raises a lot of research that creates a wide variety of framework to analyze Big Data. In this research and testing carried out an analysis of the performance of several framework that are used in analyzing the data. framework that testing were Spark, Hadoop 2, and Presto. The method used in testing is the evaluation of the time required to perform a high perfomance framework queries on certain parameters. The parameters in the test is the time spent in the low-Model query, Query middle Model, high-Model query, Query with an increase of 2 times the number of processors, and the last is a query with an increase of 4 times the number of processors. This research uses real case examples of big data analysis, data were tested in the Amazon S3 cloud storage, infrastructure analysis using cloud engine used in Qubole, testing the data analyzed is data of 1,000,000 and 3,000,000 Twitter twitter status. The end result is a performance information Spark, Hadoop 2, and Presto in high perfomance perform queries. From the information is known among the framework which has the best performance that can be a consideration in the selection of the framework to analyze the data in a variety of needs.

Kata Kunci : Big Data; Big Data Framewrok; Query; High Performance Query;Cloud Storage; Cloud Engine; Computing; Twitter