New Approach Based on Homomorphic Encryption to Secure Medical Images in Cloud Computing

Abstract

In recent years, there has been a growing demand for the adoption of cloud in healthcare to process medical data. Unfortunately, this emerging new paradigm faces several challenges, as customer data is stored on remote servers rather than on premise solutions. This is considered to be the main root cause of the major security vulnerabilities encountered in outsourced calculations. Indeed, to solve this problem we used encryption; customers should encode their sensitive data when considering adopting cloud services. One of the biggest challenges is establishing controls that completely protect secret data from insider attacks, while also supporting the computations. In addition, applying complex encryption schemes can have a negative effect on system performance.
 This study focuses in particular on homomorphic techniques and their main applications, as well as their limitations. Unlike traditional encryption methods, homomorphic schemes are used not only to reduce the security risks associated with cloud technology, but also to process cipher texts. However, current efforts in this area need to be further developed to strike the right balance between privacy risks and data utility. In this regard, we offer a hybrid approach that offers the possibility of quickly processing health records in a secure manner. Our main contribution is the proposal of a new method based on Hadoop MapReduce functions in conjunction with a multi-agent system to maintain data confidentiality. In addition, and to design intelligent distributed computing for efficient management of Virtual Machine (VM) workloads, we used a method based on the Bat Algorithm (BA).
 HIGHLIGHTS
 
 This article is about a system based on homomorphic encryption to secure medical images stored in cloud computing
 Its objective is to reduce the security risks associated with cloud technology, but also to process ciphertexts
 This system is based on the MapReduce function to benefit from parallelism during data processing and the Hadoop framework to develop and implement secure medical image processing applications in the cloud
 Hadoop and MapReduce are in conjunction with a multi-agent system to perform distributed data processing
 Our system uses a method based on the Bat (BA) algorithm to ensure efficient management of the workload of virtual machines (VMs) and thus guarantee load balancing
 
 GRAPHICAL ABSTRACT