A remote data disaster recovery system model based on Undo

The data disaster recovery system which based on fiber channel or remote data center is so costly. So the system does not appropriate to data security protection of small enterprises. The paper presents a distributed recovery method which based on remote data disaster recovery, and this method has some features including lower cost, deployment flexibility, high security etc. The data of application system is first intercepted by the Undo agent, and after the redundancy processing, finally it is stored in the distributed network storage system. In order to back up the discrete information of system, the Undo-manager sends a time signal to drive snapshoot device and sets restore points, at the same time, the captured data packets are analyzed by the system, according to the syntax and semantic, the system restores the action, system status, data and so on. When the data errors or data missing occurs, Undo-agent will take the whole into account and then drive snapshoot device to restore the checkpoints. The data which occurs after the checkpoint will replay by the verb restore module, in this way, the restoration of both coarse and fine grain and the normal running of recovered system are guaranteed.