Analysis and Implementation of Nosql Document Store Using Couchdb in Big Data Modeling

Abstract

The amount of data growth in the field of computer technology has an impact on the data produced. Data continues to grow over time, so the Big Data phenomenon appears. Big Data is data that exceeds the capacity process of a database system convention that already exists. Data with large volumes and too fast or not in accordance with the existing database architecture structure. The Relational Database Management System (RDBMS) which has become a storage and data management option has limitations to deal with Big Data issues. This research has built the NoSQL infrastructure using CouchDB. The infrastructure built consists of 3 types of storage or shard namely standalone, replica set of cluster 3 nodes and replica set of cluster 5 nodes. The infrastructure that has been built then performs performance tests that are seen from runtime, throughput, and average latency using Yahoo! Cloud Serving Benchmark (YCSB). This test is based on variations in the number of shards (1, 2 and 3 shards) and variations in data size (500,000, 1,000,000, 1,500,000, 2,000,000 records) used. The test uses the A, B, C, D, E, F, and G workloads found on YCSB, and consists of two phases namely load and run. From the test results, it can be seen that the runtime is influenced by throughput and average latency. Based on the results of the test, it was found that the performance test results were influenced by the hardware specifications used and the operations that were carried out, from the whole MongoDB better than CouchDB in read, update, insert, scan operations, but CouchDB was much better in the read-modify- operation write.