Distributed databases are gaining popularity due to advancement in technology of computer networks and due to need of the business. In distributed databases as data is located at different sites, so to get access to a particular type of data requires a query to be subdivided into subqueries and executing those subqueries at different sites. To accomplish this join operator is used. But using join sometimes incurs extra communication cost when complete relation is not required for join. In such cases to reduce the communication cost involved between two sites semi join is used. But semi join is not always useful. In this paper join operator allocation has been done dynamically by dynamically calculating selectivity factor for join and semi join for the dynamic distributed database simulated in MATLAB. This dynamic selectivity factor is given as input to the simulator built in MATLAB based on which fragment size for join operation is calculated. The simulator by using the genetic algorithm computes the minimum communication cost involved in executing the query by using combination of join and semi join.

Distributed Database Join Semi Join Selectivity Factor