Secure Sensor-based remote patient monitoring using blockchain and homomorphic encryption

Shuang Li

The development of the Internet of Medical Things (IoMT) brings opportunities but also security challenges to the healthcare field, especially in telemedicine. The growing number of cyber attacks targeting the healthcare industry increases the vulnerability of patient data to cyber threats. Once a patient’s private and personal data, which is often captured with IoT devices and processed by a third party (e.g., a cloud service provider), is leaked, it can damage the reputation of all parties concerned and result in significant financial losses. The purpose of this research is to (1) investigate the use of fully homomorphic encryption (FHE) and blockchain technology to improve data security in healthcare systems, and (2) demonstrate its feasibility in the IoMT context using a prototype platform. To evaluate the security of the proposed prototype, we simulated a scenario in which data is collected from patients, processed by a third party and made available to a doctor. Based on this scenario, we designed and implemented a prototype using a PC as a healthcare system for processing health data and a Raspberry Pi as a combination of a resource-constrained IoMT device and a mobile phone for encryption and decryption using FHE. We analyzed and evaluated (1) the performance of the proposed system using various metrics such as CPU time, memory usage, bandwidth and noise budget, and (2) its security level by measuring and controlling the FHE’s polynomial modulus degree. Our experiments show that the FHE and blockchain can be implemented on low-power IoMT devices to ensure different levels of health data security. We demonstrate that data security for transmission and processing can be achieved using FHE and blockchain with acceptable computation and memory overhead with IoMT hardware capability comparable to a Raspberry Pi. With security parameter settings representing a good tradeoff between security and performance, encryption and decryption take less than 4.6 seconds and 100MB of memory on a Raspberry Pi.

Secure Sensor-based remote patient monitoring using blockchain and homomorphic encryption

Files and links (1)

Abstract

Details

Metrics