Efficient Big Data Storage And Retrieval In Multimedia Cloud Computing Systems

Abstract

Efficient Big Data storage and retrieval is undoubtedly the biggest challenge faced by modern computing systems. In the last few years, this necessity has become more obvious due to the huge data explosion that is currently taking place on the internet and it has become a critical issue because of the wide variety of information we can retrieve from all this data.In this thesis, we study various ways of efficient storage and searching of multimedia data in cloud computing systems. Due to the plethora of information included in multimedia data, their total size does not allow us to make use of tradition storing techniques such as the use of relation database systems, or searching techniques such as xml parsing, therefore we try applying some of the more modern Big Data techniques.The approach presented in this thesis, regards, first of all, the use of a non relational (NoSQL) database for storage, and afterwards, the use of the Map Reduce programming model for extracting useful information out of multimedia data. The database we chose is MongoDB because of the high scalability potential, which plays a vital role in cloud computing systems. Moreover, using Map Reduce, we can achieve distributed data processing in a very efficient way, allowing us to execute queries in a very short time.