Privacy protection-oriented task offloading mechanism in cloud-fog collaborative assisted IoMT

Abstract

The promotion of the "smart medical" program has led to the rapid development of the Internet of Medical Things (IoMT) around the world, which has provided people with great medical convenience and especially has played an irreplaceable role in the prevention and treatment of infectious diseases. With the increasing number of IoMT users, massive data poses considerable challenges to the storage capacity and computing power of devices. Cloud computing is a good way to solve this problem. Its super large storage space and supercomputing power can not only meet the storage and computing requirements of IoMT, but also integrate and optimize resources, realize resource sharing, and save costs. According to the data privacy protection and task transfer requirements of IoMT users, combined with the collaborative structure of cloud and fog, a privacy protection awareness task offloading strategy algorithm based on federated learning was proposed, i.e., User Fairness-Task Popularity - aware task offloading (UFTP) algorithm. The UFTP algorithm uses the federated learning algorithm to train the user priority model and user preference model for the fog server. Simulation results show that, compared with the User Fairness-Aware task offloading (UFA) algorithm and Task Popularity-Aware offloading (TPA) algorithm, the proposed UFTP algorithm further considers user fairness and user preference, and the UFTP algorithm has a low task response delay.