Efficient Client-Side Deduplication of Encrypted Data with Public Auditing in Cloud Storage
The aim of the project is to reduce the cloud storage space by using the technique Deduplication of encrypted data and audit the files in cloud with the help of third party auditor to find the status of file in cloud regularly.
To reduce the volume of data, deduplication has to be performed in servers so that the storage space efficiency can be improved by removing duplicated copies. According to the research report of EMC, about 75% of the data are duplicated. In the literature, there are studies on two types of deduplication techniques. Among them, client-side deduplication has attracted the interest of researchers more than server-side deduplication due to its efficiency in computation and communication. Unfortunately, client-side deduplication has a number of problems. When clients use cloud storage services, the integrity of stored data is the most important requirement. In other words, clients want to be guaranteed about the integrity of their data in the cloud. In cloud storage services, we cannot exclude the possibility of weak cloud servers, which are vulnerable to internal and external security threats. In the case of data loss due to some incident, weak servers may try to hide the fact that they lost some data, which were entrusted by their clients. More seriously, servers delete rarely accessed users’ data in order to increase the prot. Therefore, it is a natural requirement of clients to periodically check the current state of their data.
In this paper, we design a new scheme for secure and efficient cloud storage service. The scheme supports both secure deduplication and integrity auditing in a cloud environment. In particular, the proposed scheme provides secure deduplication of encrypted data. Our scheme performs MD5 hash function for secure deduplication and integrity auditing. The proposed scheme also supports public auditing using a TPA (Third Party Auditor) to help low-powered clients. The proposed scheme satisfies all fundamental security requirements, and is more efficient than the existing schemes that are designed to support deduplication and public auditing at the same time. The main improvement in this paper is that we propose two variations to provide higher security and better performance. In the first variance, which is designed for stronger security, we assume a stronger adversary and provide a counter measure against the adversary. In the second variance, we design a technique that supports a very low-powered client and entrusts more computation to the cloud storage server in the upload procedure.