Fault tolerant distributed hash-join in relational databases
The business today are facing with immense chal- lenges due to complex applications and rapid growth in data volumes. Many of applications use data for computing statistical data to make proper decision in other applications such as machine learning or social networking. Mostly, these applica- tions assume performing sophisticated client queries with such operators as aggregation and join. State-of-the-art distributed relational databases get over these challenges. Unfortunately, distributed database management systems suffer from failures. Failures causes sophisticated queries with joining large tables are re-executed so that enormous volume of resources must be leveraged.In this paper, we introduce a new fault tolerant join algorithm for distributed RDBMS. The algorithm based on mechanisms of data replication and heartbeat messages. We compare our algorithm with traditional, unreliable distributed join algorithm. This paper demonstrates how we achieved trade-off between time required to perform tasks of failed sites and extra resources needed to carry it out.