Abstract: Data deduplication is best way for eliminating identical copies of data. This technique has been greatly used in cloud. It results in reduced storage space and network bandwidth for upload. Only one copy of file stored in cloud even though file has number of owners. This gives improved of storage space but reduces reliability. Furthermore users also stores their sensitive data on cloud. Security of this sensitive data becomes a great challenge. This research proposes new deduplication system with higher reliability. This technique is used to save storage space and bandwidth under cloud. In this research data chunks are stored on multiple cloud servers. Data deduplication has two types one is file level deduplication and other is block level deduplication. The new deduplication system also maintains direct communication between deduplication System’s users and owner of the file. Users can directly communicate with owner and request for file. Owner of the file send the respective file to the user in secure way. The deduplication system is mainly used in educational institutes and industry. The deterministic secret sharing scheme is used to achieve security requirements of data privacy and tag consistency. This research incurred small overhead in realistic environments.
Keywords: Deduplication, cryptography, distributed storage system, reliability, security.