Low Overhead Incremental Checkpointing and Rollback Recovery Scheme on Windows Operating System

Implementation of a low overhead incremental checkpointing and rollback recovery scheme that consists of incremental checkpointing combines copy-on-write technique and optimal checkpointing interval is addressed in this article. The checkpointing permits to save process state periodically during failure-free execution, and the recovery scheme maintains to normally execute the task when failure occurs in a PC-based computer-controlled system employed with Windows Operating System. Excess size of capturing state and arbitrary checkpointing results in either performance degradation or expensive recovery cost. For the objective of minimizing overhead, the checkpointing and recovery scheme is designed of Win32 API interception associated with incremental checkpointing and copy-on-write technique. Instead of saving entire process space, it only needs to save the modified pages and uses buffer to save state temporarily in the process of checkpointing so that the checkpointing overhead is reduced. While system is encountered with failure, the minimum expected time of the total overhead to complete a task is calculated by using probability to find the optimal checkpointing interval. From simulation results, the proposed checkpointing and rollback recovery scheme not only enhances the capability of the normal task executing but also reduces the overhead of checkpointing and recovery.