NEMESIS: No-Regret E-Health User Experience in Multi-Access Edge Computing Systems

The rapid growth of data and computing needs in the Internet of Medical Things (IoMT) necessitates efficient mechanisms for optimizing the resource management in e-health applications. This paper presents the NEMESIS framework, which enables the users to determine their optimal Multi-Access Edge Computing (MEC) server selection and data offloading strategies by considering the reliability of the MEC servers based on individual interactions and shared user experiences. A comprehensive system model is introduced that defines the users’ interactions, the data offloading processes, and the impact of various IoMT devices, along with a novel utility function that evaluates the tradeoffs in the MEC server selection and task offloading. Additionally, a reliability model is proposed that incorporates the direct user interactions and their peers evaluations of the MEC servers’ computing services, while a regret learning mechanism is designed to optimize the users’ strategies under varying information scenarios. The results demonstrate that the NEMESIS framework operates efficiently in real-time and outperforms state-of-the-art scheduling and offloading schemes in terms of latency and energy consumption.