Optimization-Assisted Secure Authentication for Medical Devices in the Internet of Medical Things (IoMT)

Abstract

The increase in medical devices leads to the creation of the IoMT (Internet of Medical Things) environment where the authentication process is one of the most important features that is a must in terms of cybersecurity and a bottleneck from the performance point of view. Medical sensors, wearable devices, bedside devices, and other medical devices can have limited computing power, battery power, memory, and wireless connection reliability. Therefore, the authentication process should not use too much resources so that latency and energy losses could be avoided, offering_inadequate_security against impersonation, replay, man-in-the-middle, and other attacks.

The paper provides a realistic benchmark for evaluating the authentication framework’s performance. A benchmarked dataset consists of 12000 authentication requests, six device classes, and five attack classes with benign traffic. The framework’s performance is compared to ECC-Baseline, LightECC, PSO-ECC, and the proposed hybrid policy. According to the experiments, the suggested framework can achieve 12.83 ms of mean authentication latency, 0.46 mJ of energy consumption per authentication request, and 98.06% attack detection rate. The results are promising, yet the experiments are made in a simulated environment, not on real hardware. Moreover, the results are theoretical; the actual attack surface depends on the implementation details and side-channel exposure.

The paper provides a comprehensive description of the proposed framework along with a formalized optimization function, realistic environment, attack surface, dataset generator, algorithm, and security evaluation. It is possible to utilize the information provided in this paper to test the framework on real hardware and publicly available datasets.