Design and analysis of cryptographic schemes for secure data processing in distributed storage systems

Abstract

The advancement in internet of things (IoT) and its adoption in every aspect of life has triggered many applications which led to an exponential increase in the global data generation rate. Distributed storage systems (DSS) offer a promising solution to the ever growing need for a flexible and scalable storage resource. In order to ensure data availability in DSS, data is redundantly stored among several inexpensive storage devices. Among the various methods for incorporating data redundancy, regenerating codes (RC) form an advanced coding technique which provide data availability with an optimal trade off between storage overhead and repair bandwidth overhead. Distributed storage architecture is widely adopted by cloud storage systems and wireless sensor networks (WSNs). In WSN the data is freely accessible over wireless channels; while in cloud platform, the data is stored in untrusted third-party cloud servers. Therefore ensuring data security and reliability is paramount in these DSS applications. Traditional encryption schemes protect data confidentiality; but they cannot support processing operations in encrypted domain (ED). In order to support various data processing services in DSS applications without sacrificing security, the encryption scheme should have homomorphic properties. Also, ensuring data reliability and integrity in DSS is a challenging task because the data in the nodes may get corrupted through channel errors, storage/memory errors or tampering by intentional adversaries. This research work aims to design and develop methods to fortify data reliability, confidentiality and integrity in distributed storage applications while supporting various data processing operations. newline