Adaptive staircase multiple failure detector for parallel and distributed image processing

Failure of multiple nodes detection is a complex task in parallel and collaborative image processing on a cluster. Adaptive heartbeat detectors with complex prediction system are not effective due to high overheads as image processing itself having high time complexity and thus make such detectors unrealistic. In this paper an improved algorithm is proposed which combine stair case heartbeat mechanism with adaptive automatic time predication mechanism for such image processing application. All nodes are broadcasting with “I am ok” message by setting the time initially min time. All nodes who answer are kept in a first working node list. Again time is increased in stair case manner for only those nodes who had not answered. This is repeated till maximum time is reached. After this, node is declared as a faulty node if it does not answer. It improves the accuracy of the multiple failure detectors since this algorithm relates the time of different node as per the load and traffic by gradually determined stair case intervals. It is simple since no complex and time-consuming time prediction mechanism is incorporated. Simulation Results presented validate the effectiveness of the proposed algorithm.