Analysis of Edge Intelligent Frameworks and their Security Issues

Abstract

Edge Intelligence has become increasingly popular and has already made its place to increase the overall system performance by reducing the burden of the cloud and the network. In edge intelligent frameworks, a massive amount of data generated are not provided to the central cloud, and data analysis is carried out at the edge. Edge intelligence IoT environments comprise heterogeneous devices that communicate over the network, making it essential to protect the data and users’ information. Through these edge frameworks, numerous users and devices take part in communication where the exchange of sensitive data occurs. Therefore, security in such frameworks is crucial and a key challenge for reliable communication. This paper performs an analysis of popular AI/ML applications toward edge intelligence focusing on highlighting the critical security and privacy concerns desired in such systems. After a thorough investigation, we show that although several promising edge intelligent frameworks have been developed to address energy and performance issues, they do not consider the security and privacy of the data as the researchers are more focused on the performance predicaments.