Protecting Medical Data in Cloud Storage Using Fault-Tolerance Mechanism

Given the fact that cloud computing offers cost-efficient storage systems, medical organizations are more interested in using this alternative solution to safeguard their patients' data. Equally interestingly, users are charged based typically on the amount of occupied storage space. Basically, this concept is meant to cut costs and improve the quality of healthcare services. Consequently, implementing cloud storage would help clients to manage their data efficiently. Besides, it allows users to outsource the storage process by using virtual storage systems instead of local ones. Despite its significant impact in healthcare domain, adopting this paradigm to save medical data on remote servers poses serious challenges, especially security risks. Currently, various cryptographic techniques have been used to ensure data confidentiality and to avoid data disclosure. Globally, this model uses traditional cryptosystems such as AES, RSA to address security issues in cloud storage. As far as we know, there are only a few works in literature that deal with availability and data recovery in cloud computing. In general, the classical approach which is based on backup or replication is not suitable for cloud environment due to the highly dynamic nature of this model. The intent of this work is to enhance the reliability of cloud storage in order to meet security requirements. In this study, we propose a novel method based on Shamir's Secret Share Scheme and multi-cloud concept to avoid data loss and unauthorized access. More precisely, this technique seeks to divide consumers' data into several portions using Shamir's Secret Share to prevent privacy disclosure. Based on these considerations, we store these created portions in different nodes to minimize security risks, particularly internal attacks. To sum up, this method is designed to ensure fault-tolerance, which is the main subject of this study. In fact, we need just certain shares to reconstruct the secret data rather than using all parts. The experimental results are in accordance with the theoretical assumptions behind this model, and hence, confirm that the proposed framework provides necessary measures for preventing data loss in cloud storage.