An Application of Bayesian Theorem for Optimal Replica Placement in Data Grids

T. Amjad and M. Alam (Pakistan)


Grid Computing, Data Grid, Data Replication, Replica Placement.


Data replication is a well known solution to provide high data availability in data grid environment. Optimal placement of replica is very important to achieve the maximum possible gains from replication. There is a wide range of techniques that are available for Replica Placement in data grid environment. Any replication algorithm must answer some very critical questions like; when to create a replica? where to place it? This involves decision making in replication. Bayesian Theorem is a well known statistical method which is currently being used in many disciplines of sciences. In this research Bayesian Theorem is applied in decision making for the optimal Placement of a replica, keeping in view the cost of replication, distance, available bandwidth, and available storage resources to ensure maximum availability of data. It can be applied in the scenarios where decisions have to be made under uncertain conditions. Also it runs in a relatively cost effective way in real time environments. The proposed replication algorithm has a run time complexity of O(n), which shows that presented approach will improve availability, response time, and scalability in data grid environment.

Important Links:

Go Back