Research And Implementation Of Block-level Continuos Data Protection Service Based On Group

Nowadays, enterprises have been than ever more dependent on data. How to back updata and recovery from disaster become the research hot spot of storage industry. However,traditional data protection technology more and more has become unable to meet theneeds for a much better Recovery Time Objective (RTO) and Recovery Point Objective(RPO), then Continuous Data Protection (CDP) technology emerge and becomeincreasingly popular.Most existing implementations of CDP are either storing data to the dedicatedrepository, or relying on the protected host using remote storage space. There is high cost,less efficient, inconvenient expansion and other issues in either of the program. In thispaper, a CDP service based on group model is proposed. Efficient block data access isprovided via group copy management. B+tree and index page management is used tosearch metadata quickly. The shortest recovery path is found by branch file together withincremental and full recovery algorithm. Also，based on synergy recovery, the recoveryrequest can be finished quickly with minimal resource. By scheduling the allocator ofresources, we provide responsive, fair CDP service. Taking advantages of hosts and LAN,CDP dramatically improves the efficiency and scalability while reducing costs.It can be seen that the system can support multiple users simultaneously using CDPservice with every user more than8checkpoints. Lower service latency(23%) guarantee ofequitable and timely response service. The structure based on group reflects higherperformance on recovery time compared with HUST-CDP.