The role of mobile agents in increasing speed up repair failure in the mobile database systems

The mobility of the hosts in mobile database system makes it difficult to access the log and take checkpoint, while it is being repair failured. So repair failure is a very complicated process in mobile database. All the processes in one host must take checkpoint in order to ensure global checkpoint consistency. Some methods results in putting and storing checkpoints and sending a huge number of messages among processes. In this article, we used the mobile agents in order to improve the time of checkpoint and as a result, the total time of repair failure will be decreased. We aim that it´s possible via collecting the information of processes involved in transaction and then putting the information in mobile agent moving with mobile host. So, by having the list of involved processes, just the request of placing checkpoint is given to them and forces a minimum number of processes to take checkpoint. This paper represents an unblocked coordinated low latency checkpoint in order to facilitate the repair failure based on mobile agent´s characteristics, processes are unblocked but a small temporary log, which is always connected to the main log, is used for saving the messages exchanged from the time of request for checkpoint by the starting process till the time of the stabilization or the cancellation of that checkpoint. We used an environment multiagent for distributed system-JADE- to implement.