Perbandingan Metode Mapreduce Berbasis Single Node Hadoop Pada Aplikasi Word Count

Abstract

 In the context of Big Data processing, Hadoop MapReduce serves as a framework used to develop software and process large-scale data in parallel. Word Count is a type of task used to count the occurrences of unique words in a text file. Considering processing time is crusial in adhering to standards of Big Data Processing. The conducted research involved the processing of text files using the MapReduce method on the Hadoop Distributed File System (HDFS) using a single node, comparing the results of Word Count processing with and without the use of MapReduce. The research findings indicate that the implementation of Word Count without using MapReduce offers better speed in processing Indonesian language text data on a Hadoop single node. Additionally, the comparison of processing time between the Word Count program with Hadoop MapReduce and the Word Count program without MapReduce shows that the latter has faster processing time. A significant reduction in processing time, up to 95% for a 5 MB file size, can be achieved by using the Word Count method without MapReduce. However, the level of reduction decreases with increasing file size.