Efficient Resource Management Framework for Critical Healthcare Applications in Integrated Edge Fog Cloud Environments using Blockchain based Federated Learning Methods

Abstract

A critical system in real-time generally manages any environment by receiving the input, processing it, and producing output data while meeting specific time constraints. In such systems, deadlines must always be satisfied irrespective of the load of the system. As the number of real-time applications that need smart connections among each other increases, the Internet of Thing (IoT) challenges will also increase. The new evolving technology that may help in time-critical realtime IoT-based systems is the use of Edge/Fog. To make use of the advantages of edge/fog computing for real-time critical applications, we propose an integrated system of edge, fog, and cloud computing environments for the healthcare sector. By eliminating irrelevant data at the Edge/Fog nodes, the proposed architecture saves time as well as minimizes the amount of data that must be transferred to the cloud. Resource management is one of the primary challenges in a real-time environment since the demand for resources grows dynamically and needs rapid provisioning and processing. Resource management techniques in edge/fog are challenging because the modules of analytic applications are moved to each edge device to minimize the delay and avoid congestion. This research focuses on developing efficient Edge-Fog-Cloud-based resource management for medical applications to newlineavoid delay and improve the efficiency of smart healthcare systems. To maximize the utilization of these resources and improve the efficiency of applications, efficient newlinescheduling and resource allocation are required. To avoid over-provisioning and under-provisioning, dynamic resource provisioning, an effective method of preparing newlineresources based on changes in the workload of IoT applications, is required. Due to the vast solution space, scheduling in fog/edge computing is classified as an NP-hard problem, which means it takes a long time to discover an optimal solution. No algorithms can handle these issues in polynomial time and yield optimal results. Finding a sub-optimal...