东南大学学报（英文版）2019，Vol.35Issue(3)：325-331,7.DOI:10.3969/j.issn.1003-7985.2019.03.008
多对一多服务器蜂窝边缘计算延时优化资源调度
Delay-performance optimization resource scheduling in many-to-one multi-server cellular edge computing systems
摘要
Abstract
To further reduce the delay in cellular edge computing systems,a new type of resource scheduling algorithm is proposed.Without assuming the knowledge of the statistics of user task arrival traffic,the analytical formulae of the communication and computing queueing delays in manyto-one multi-server cellular edge computing systems are derived by using the arriving curve and leaving curve.Based on the analytical formulae,an optimization problem of delay minimization is directly formulated,and then a novel scheduling algorithm is designed.The delay performance of the proposed algorithm is evaluated via simulation experiments.Under the considered simulation parameters,the proposed algorithm can achieve 12％ less total delay,as compared to the traditional algorithms.System parameters including the weight,the amount of computing resources provided by servers,and the average user task arrival rate have impact on the percentage of delay reduction.Therefore,compared with the queue length optimization based traditional scheduling algorithms,the proposed delay optimization-based scheduling algorithm can further reduce delay.关键词
蜂窝系统/延时/边缘计算/资源调度Key words
cellular system/delay/edge computing/resource scheduling分类
信息技术与安全科学引用本文复制引用
杜鹏,巴特尔,张源..多对一多服务器蜂窝边缘计算延时优化资源调度[J].东南大学学报（英文版）,2019,35(3):325-331,7.基金项目
The National Natural Science Foundation of China (No.61571111). （No.61571111）
