An automate failure recovery for synchronous distributed database system

Abstract

Periodically, researchers have been sharing their constant attempts to improve the existing methods for data replication in distributed database system. The main goal is to work for an efficient distributed environment. An efficient environment may handle huge amount of data and preserve data availability. The occasionally failures in distributed systems will affect the end results, such as data loss, income loss etc. Thus, to prevent the data loss and guarantee the continuity of the business, many organizations have applied disaster recovery solutions in their system. One of the widely used is database replication, because it guarantees data safety and availability. However, disaster still can occur in database replication. Hence, an automatic failure recovery technique called distributed database replication with fault tolerance (DDR-FT) has been proposed in this research. DDR-FT uses heartbeat message for node monitoring. Subsequently, a foundation of binary vote assignment for fragmented database (BVAFD) replication technique has been used. In DDR-FT, the data nodes are continuously monitored while auto reconfiguring for automatic failure recovery. From the conducted experiments, it is proved that DDR-FT can preserve system availability. It shows that DDR-FT technique provides a convenient approach to system availability for distributed database replication in real time environment.