Homomorphic cryptosystem-based secure data processing model for edge-assisted IoT healthcare systems

Abstract

The substantial acceptance of Internet-of-Things (IoT) in digital healthcare systems necessitates secure data processing before storing it on a cloud server. Since resource-limited IoT devices engender complexity in secure processing, edge computing technology is adopted. Edge-assisted IoT locates edge nodes in proximity to user devices enabling computation offloading and seamless interaction between IoT applications and cloud servers. Integration of these entities results in severe security and privacy issues due to large amount of sensitive information transmitted over vulnerable communication channels. As such, information protection and data integrity in edge-assisted IoT have become critical factors for improvement. Several data processing models have been proposed regarding the above-stated concerns. However, under a powerful adversarial scenario, traditional solutions lack in preserving robust security and privacy. This motivates us to address critical challenges of communication security and privacy-preserving computation offloading in edge-assisted IoT systems. In this article, we design a homomorphic cryptosystem-based secure data processing model (HC-SDPM), enabling secure data collection and aggregation offloading on edge nodes for our hypothesized edge-assisted IoT healthcare system. Paillier homomorphic encryption and certificateless linear homomorphic ID-based signatures are utilized to achieve end-to-end confidentiality and data integrity during transmission and aggregation. Moreover, HC-SDPM involves a semi-trusted authority to enable key-escrow resilience. The security analysis of HC-SDPM assures signatures unforgeability in the random oracle model under computational Diffie–Hellman assumptions. Compared with similar modeled schemes, HC-SDPM demonstrates a relative advantage in computational overhead, communication overhead, and storage complexity.