Enhancing the Availability and Partition Tolerance of NoSQL Databases

Abstract

Storage management is one of the most popular challenging tasks in cloud computing data storage. NoSQL became widely known data stores in distributed data storage for two reasons: (1) high until almost ultimate scalability and (2) low administration overhead. In these days, the growth of data is extremely fast and thus scalable distributed data store is essential for the storage of large volumes of data (petabytes, terabytes of data). NoSQL databases are the most suitable solutions for the above situation as well as in cloud computing. Distributed databases have three properties that are commonly desired are: CAP (consistency, availability, partition tolerance). It would be ideal if distributed databases can deliver all three guarantees. However, providing all three attributes in NoSQL is still a challenge. This paper intends to examine how NoSQL works and to enchance the availability ,by using replication mechanism and latency reducing mechanism, in NoSQL databases that are suitable for cloud computing environment. Moreover, the paper also presents the enhancement on partition tolerance in NoSQL data-stores by using the symmetric replication mechanism in P2P chord ring.