Checkpointing of parallel applications in a grid environment

Jobs in Grid workflows are exposed to different types of failure. It is important to develop fault tolerant mechanisms to ensure a good level of reliability during the execution of Grid jobs. While checkpointing is the most common method to achieve fault tolerance, there still is a lot of work to be done to improve the efficiency of the mechanism. The paper gives an overview of a checkpoint solution for checkpointing parallel applications executed on multiple sites in the Grid environment. The checkpointing mechanism is an improvement of the PGRADE checkpointing solution