Research Article
A Balanced Scheduling Technique for Distributed Inference Resources Based on Edge Computing
@INPROCEEDINGS{10.1007/978-3-032-00303-4_10, author={Ruiwu Chen and Xuxia Wang}, title={A Balanced Scheduling Technique for Distributed Inference Resources Based on Edge Computing}, proceedings={Advanced Hybrid Information Processing. 8th International Conference, ADHIP 2024, Jiaxing, China, September 20--22, 2024, Proceedings, Part III}, proceedings_a={ADHIP}, year={2026}, month={11}, keywords={Edge computing Distributed reasoning resources Resource scheduling Balanced scheduling Resource matching}, doi={10.1007/978-3-032-00303-4_10} }- Ruiwu Chen
Xuxia Wang
Year: 2026
A Balanced Scheduling Technique for Distributed Inference Resources Based on Edge Computing
ADHIP
Springer
DOI: 10.1007/978-3-032-00303-4_10
Abstract
The load indicators of CPU utilization, memory utilization, disk space utilization, and network bandwidth utilization increase the complexity of resource scheduling. In order to achieve the goal of balanced scheduling of distributed reasoning resources, a new balanced scheduling technology is proposed by using edge computing. First, evaluate the load. Monitor and evaluate the load of each edge node in real time according to four indicators, so that the most suitable node can be selected to perform tasks when scheduling tasks. Secondly, according to the characteristics of the task and the real-time status of the node, the task is matched with the distributed reasoning resources. Finally, based on the matched resources, the edge computing is used to design a balanced scheduling algorithm for distributed reasoning resources, and the balanced scheduling of distributed reasoning resources is achieved by executing the algorithm. The test results show that the application of this technology can significantly improve resource utilization, achieve better load balancing, and make resource allocation more balanced.