An Online Two-Phase Workload Management Scheme for Collaborative Edge Computing System

The rapid expansion of Internet of Things (IoT) de-vices poses substantial challenges to the Quality of Service (QoS) provided by edge servers, which are often constrained by limited resources, particularly when handling numerous time-sensitive tasks. To address these challenges cost-effectively, collaborative edge computing has been introduced. This approach enhances system performance by enabling edge servers to offload tasks to neighboring edge servers or remote cloud servers, thereby alleviating the computational burden. However, existing research often falls short in providing robust long-term solutions for real-time online scenarios where tasks arrive unpredictably, lacking prior information. To this end, this paper explores the problem of online workload management in collaborative edge computing systems, aiming to improve the long-term utility of edge servers. We formulate the problem as a non-linear optimization problem. To solve this problem efficiently within polynomial time, we intro-duce an online two-phase workload management scheme called OTWMS. This scheme breaks down the workload management problem into several distributed and parallelizable local resource allocation problems and a centralized task migration matching problem. Through evaluation experiments, we demonstrate that the proposed scheme surpasses several benchmarks in terms of long-term utility and service performance optimization.