Survey on Data Replication in DRTDBS: Taxonomy and State of the art replication technique enhances the performance of the distributed real time database system (DRTDBS) via sharing the load among available replicas, decreasing the bandwidth utilization, locally processing the real time transaction, and timely providing the data to the admitted real time transactions (RTTs). This technique has been broadly studied phenomena in the replicated distributed real time database system (RDRTDBS) in which data replicas are spread over multiple locations. The major challenging issue in this technique is to solve 4W-H. This paper presents a detailed survey of data replication in DRTDBS. Specifically, our survey includes replication strategies, replication protocols, QoS, dependency relationship, security, fault tolerance, and scalability. Based on the current state of the art, we have further classified replication strategies into three dimensions: (i) static vs. dynamic, (ii) active vs. passive, and (iii) synchronous vs. asynchronous. Similarly, the replication protocol is also classified two times. In the first classification, replication protocols are classified into three dimensions: (i) certification-based protocol, (ii) weak voting protocols, and (iii) deterministic protocols. In the second classification, replication protocol is classified into two dimensions: (i) kernel-based replication and (ii) middleware-based replication. A tabular representation is presented to facilitate the future comparison of data replication methods in RDRTDBS. This survey provides a technical direction for researchers on RDRTDBS systematically and the potential future aspects have been discussed.
Volume 12 | 01-Special Issue
Pages: 925-943
DOI: 10.5373/JARDCS/V12SP1/20201144