Survey of NoSQL Database Engines for Big Data

Cloud computing is a paradigm shift that provides computing over Internet. With growing outreach of Internet in the lives of people, everyday large volume of data is generated from different sources such as cellphones, electronic gadgets, e-commerce transactions, social media, and sensors. Eventually, the size of generated data is so large that it is also referred as Big Data. Companies harvesting business opportunities in digital world need to invest their budget and time to scale their IT infrastructure for the expansion of their businesses. The traditional relational databases have limitations in scaling for large Internet scale distributed systems. To store rapidly expanding high volume Big Data efficiently, NoSQL data stores have been developed as an alternative solution to the relational databases. The purpose of this thesis is to provide a holistic overview of different NoSQL data stores. We cover different fundamental principles supporting the NoSQL data store development. Many NoSQL data stores have specific and exclusive features and properties. They also differ in their architecture, data model, storage system, and fault tolerance abilities. This thesis describes different aspects of few NoSQL data stores in detail. The thesis also covers the experiments to evaluate and compare the performance of different NoSQL data stores on a distributed cluster. In the scope of this thesis, HBase, Cassandra, MongoDB, and Riak are four NoSQL data stores selected for the benchmarking experiments