Performance Analysis of NoSQL Databases with YCSB

• Main Authors: Tsai,Chi-Hao, 蔡淇浩
• Other Authors: 吳東光
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/63193383149173014758

碩士 === 國立彰化師範大學 === 資訊管理學系所 === 104 === Nowadays, NoSQL databases are becoming to be the mainstream database management systems (DBMSs) in the Big Data era. In this thesis, we select the top two most popular NoSQL DBMSs according to the ranking made by the American database knowledge website, the DB-engine (http://db-engines.com/), in January 2016, to be our research subjects. The two are MongoDB and Cassandra database systems, which fall into the document-oriented and wide column store database categories, respectively. Our research focuses on the performance analysis of the above mentioned two NoSQL databases. YCSB (Yahoo! Cloud Serving Benchmark) is used as the benchmarking tool for the evaluation. The study will consider several aspects, which include the ease of deployment and scaling, the improvement in performances in terms of scaling, and the storage usage and the ease in managements. In our experiments, we use work load configurations that are combinations of a variety of operations, such as read, write, modification, fixing and scan. Moreover, we also consider various loading scenarios range from a single thread to multiple-thread executions of YCSB to investigate thoroughly the performance of these two NoSQL DBMSs.