BEC: A Reliable and Efficient Mechanism for Cloud Storage Service

In a cloud storage system, the hardware equipments storing the contents are located in many places and the capabilities of them are various. These realities draw a big challenge to providing constant and efficient service to the storage system's users. In this paper, intrigued from the features of erasure coding, we design a content storage and delivery mechanism which can tolerate cloud servers' failures, so that ensure the reliability of the cloud storage system while distributing contents efficiently. In this mechanism, the users can still retrieve an adequate number of coded packets to recover the original packets after a fixed number of servers collapse. Afterwards, through experiments we compare our mechanism to the erasure coding method and the replication redundancy strategy, and validate that our technology takes advantages in fault tolerance and repair efficiency.