A Profit-Driven Resource Management Scheme for Collaborative Edge-Cloud Computing

In edge computing, the disparity between static available resources and dynamic computational workloads can cause edge servers to become either overburdened or under-utilized. This imbalance, characterized by either insufficient or idle resources, leads to reduced quality of service and decreased profitability, especially when handling a large number of latency-sensitive tasks. To address this issue, collaborative edge-cloud computing, which allows edge servers to execute all complex tasks with the help of the cloud, has been proposed as a potential solution. However, few studies have focused on optimizing profit through efficient resource management. To this end, we investigate how to achieve higher profit in the collaborative edge-cloud computing system by efficiently sharing, expanding and purchasing resources. In this paper, we model this resource management problem as a profit optimization problem, aiming to determine the optimal strategies for resource expanding, sharing, and real-time purchasing to maximize total profits. To solve this optimization problem effectively, we first analyze the distribution of computational workloads and the relationships among shared resources, expanded resources, and the minimum expected purchasing costs. Based on this analysis, we then design a profit-driven resource management scheme (PDRM). It first maximizes expected profit through an efficient resource expanding and sharing approach, then minimizes purchasing costs by employing a real-time resource purchasing strategy, ultimately optimizing overall profit. Finally, through designed experiments, we demonstrate that our proposed profit-driven resource management scheme can enhance total profits while satisfying the low-latency requirements of computation tasks.