Capturing Anomalies of Cassandra Performance with Increase in Data Volume: A NoSQL Analytical Approach

Abstract

NoSQL database technology has been doing rounds since the early 1990s, but it was the exponential growth of internet and the rise of web applications that lead to a dynamic surge in the popularity of NoSQL databases. The BigTable research by Google (2006) and the Dynamo research by Amazon (2007) paved the way for databases which could develop with agility and operate at any scale. Cassandra and MongoDB have emerged as the two most widely used NoSQL database and hence either of the two is preferred depending on the data problem user is attempting to solve. This paper describes the underlying principles as well as the differences between both the databases. We focus on showing the anomaly in performance of Cassandra as the data volume increases and at the same time we compare its performance with that of MongoDB. We establish how important factor is data volume in choosing either of the databases for an application. Extensive experiments have been carried out to scale the performance in terms of anomaly similarities, and the future scope is pinpointed.