A secure regenerating code for the fault-tolerant of distributed networked storage

Regenerating code is a fault-tolerant technique based on the network coding. However, most of the researchers just focus on the construction of the codes, ignoring the security of the codes in application, especially the security in an extremely open environment such as cloud computing and cloud storage. A secure regenerating code, called SRCS is proposed in this paper to solve the fault-tolerant of distributed networked storage. SRCS achieves the security of the regenerating code using the threshold public-key encryption with low redundancy and high efficiency. In the storage system based on the SRCS, the data owner shares his private key to a set of key servers and uses the storage servers to stores the data. Even if the attacker compromises all storage servers, he can't get the data. It shows that SRCS is secure in the semi-adaptive model using decisional BDHE assumption.

I. Introduction

Regenerating code^[1] is a fault-tolerant technique based on the network coding. Assume that there are storage servers in the networked storage system. In Regenerating code, the messages will be encoded into data blocks and stored into the storage servers. If there are more than storage nodes are active the original data can be regenerated. Compared with the traditional fault-tolerant techniques based on erasure code, network coding can effectively reduce the transmission bandwidth cost during the repair process. Hu^[2] and Rashmi^[3] proposed two optimal models of the RC code which can reduce more transmission bandwidth cost. Jun Li^[4] proposed a optimal regenerating process for the RC code with a higher efficiency.