A simulation tool for a large-scale NoSQL database

Abstract

The amount of data on the Web and the number of users accessing those data drastically grows every year. Web-based services have to be able to receive and process those requirements within the shortest possible time. Additionally, they have to manage sudden and unpredictable peaks of user requirements for accessing the data. To tackle these problems, NoSQL databases have been designed to provide scalability and good performance. However, it is important to know in advance the number of resources (i.e. computers, network capacity) to efficiently deploy a database application. In this work, we propose a simulator for a NoSQL database named MongoDB. Our simulator aims to detect saturation levels of the resources to ensure an efficient execution of MongoDB upon different users demands and upon failures of resources.