Grid computing is a coordinated resource sharing and solving the problems in organisations which are dynamic and virtual in nature. Many grid applications will be running in environments where interaction faults more commonly occur between diverse grid nodes. As resources may also be used outside of organisational boundaries, it becomes iteratively difficult to guarantee that a resource being used is not malicious one. Because of the diverse faults and failure conditions developing, deploying, and executing long running applications over the grid remains a challenge. Hence, fault tolerance is a primary factor for grid computing. The prototype system here is designed using agents to provide service replication, reactivation and avoids the single point of failure in a simulated computational grid environment. The agents and the workflows are provided by a common software platform called WADE. Moreover, workflows in the computational grid can also be generated using WOLF which is embedded in the WADE platform itself. Copyright © 2013 Inderscience Enterprises Ltd.