Optimization Model and Simulation Platform for Task Scheduling Between Edge Servers in Edge Computing Systems

Edge computing has emerged as a popular paradigm to meet the growing demands of time-sensitive and compute-intensive tasks. In edge computing systems, the problem of task load imbalance among edge servers often leads to high response latency, which negatively affects the performance of individual server. Therefore, a well-designed scheduling strategy for edge task plays a crucial role in improving the overall system performance. There is some existing work on solving the task load imbalance problem by scheduling between devices and edges, or between edges and cloud servers. Unlike the existing work, in this paper we perform task scheduling between edge servers, using low load edges to assist high load edges for task processing. However, due to the multiple tasks that are processed on each edge server, the edge server state is complex, scheduling tasks between edge servers is a difficult problem. To address this problem, we consider modeling each type of task to be processed on edge servers as a queue, and formulate the optimization model with the constraint that the sum of edge service time within a time slot is less than the slot length. And then we model the task scheduling problem between edge servers as an optimization model. Since we need to validate our proposed model through the simulation platform, and the existing simulation platform does not support task scheduling between edge servers, we design a simulation platform with edge scheduling capabilities to validate the effectiveness of our proposed model. Through simulation experiments, it is verified that the edge task scheduling policy obtained by our optimization model can well improve the performance of the edge computing system. Moreover, since our simulation platform is modular developed based on Python programming, it can be easily applied to other optimization models. The code for our optimization model and simulation platform is available at https://github.com/AtticusZhang/PyEdgeSim.