Fault tolerant mechanism in grid based on Backup Node

摘要

Grid is a very efficient technology to performing heavy processing with distributed resources. These resources at geographically distributed widely and without central control unit. One of the important challenges in grid is fault tolerant, because the grid resources in this environment are not reliable. And to avoid duplication with processing done by a resource, the Check Pointing mechanism has been proposed. The state of application at particular point of time can be store and in case of failure the status information can restore on new node and computation can continue. Check Pointe storage location is a major challenge in the fault tolerant techniques. In this paper we presented a new method by using a backup node. Adding a component to GRAM to improve efficiency and fault tolerant. With clear Backup Node the Check point information can save on it. This method increases the efficiency and system reliability.