Vers un nouveau modèle de stockage et d‘accès aux données dans les Big Data et les Cloud Computing

Abstract

The technological revolution integrating multiple information sources and extension of computer science in different sectors led to the explosion of the data quantities, which reflects the scaling of volumes, numbers and types. These massive increases have resulted in the development of new locations techniques and access to data. The final steps in this evolution have emerged new technologies : Cloud Computing and Big Data. Big Data is a set of technologies based on NoSQL databases allowing scalability of volumes, numbers and types of data. The important companies in the IT sector find these NoSQL systems, new solutions to respond to scalability needs. Multiple open-source and proprietary models of NoSQL are available on the market. In the first contribution, we develop a comparative study about the performance of six solutions widely employed MongoDB, Couchbase, Cassandra, HBase, Redis and OrientDB, using the YCSB tool. The latter being a very powerful test tool, used in multiple evaluation works of NoSQL databases. The purpose is to provide assistance and support to actors interested of Big Data and Cloud Computing for eventual decisions for the choice of solutions to be adopted. Our second contribution consists in proposing internal revisions in the architecture of the reference implementation of storage clouds and Big Data, namely HDFS of Hadoop. This latter is based on the separation of metadata to data that consists in the centralization and isolation of the metadata of storage servers. We propose an approach to improve the service metadata for Hadoop to maintain consistency without much compromising performance and scalability of metadata by suggesting a mixed solution between centralization and distribution of metadata to enhance the performance and scalability of the model.