云环境下协同作业的密码服务优化调度算法  

Cryptographic service optimization scheduling algorithm for collaborative jobs in cloud environment

作　　者：曹晓刚 李凤华[1,2,3] 耿魁 李子孚[1,3] 寇文龙 CAO Xiaogang;LI Fenghua;GENG Kui;LI Zifu;KOU Wenlong(Institute of Information Engineering,Chinese Academy of Sciences,Beijing 100085,China;School of Cyber Security,University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Cyberspace Security Defense,Beijing 100085,China)

机构地区：[1]中国科学院信息工程研究所,北京100085 [2]中国科学院大学网络空间安全学院,北京100049 [3]网络空间安全防御重点实验室,北京100085

出　　处：《通信学报》2024年第7期84-100,共17页Journal on Communications

基　　金：国家重点研发计划基金资助项目(No.2021YFB3100300);国家自然科学基金资助项目(No.62202463);中国科学院青年创新促进会人才基金资助项目(No.2023168)。

摘　　要：针对云环境下密码按需服务中多个计算作业协同服务的需求,提出了多密码作业协同服务的调度算法,能够充分应对密码算法种类多、并发需求高、作业随机交叉和作业负载瞬时激增等云环境下的新挑战。考虑每个密码计算作业之间的依赖关系、密码作业的完成时间需求以及密码计算单元的最大算力,以最小化能耗、迁移成本和瞬时激增负载的适应度为优化目标,将多密码作业协同服务调度问题建模为多目标优化的作业流调度问题,并提出“选择-排序”两阶段调度算法,在选择阶段,采用改进NSGA-Ⅲ算法为密码计算作业选择合适的计算单元,在排序阶段,根据作业紧迫程度决定执行顺序。仿真结果表明,所提调度算法在能耗、迁移成本和对瞬时激增的作业负载的适应度方面优于传统调度算法。In response to the demand for collaborative computation of multi-cryptographic jobs in cryptographic on-demand services within a cloud environment,a multi-cryptographic job collaborative scheduling algorithm was proposed.This algorithm effectively addressed new challenges in cloud environments,such as a variety of cryptographic algorithm types,high concurrency demands,random cross-job interactions,and sudden increases in workloads.Considering the dependencies among jobs,makespan of jobs and computational power of computing units,the scheduling problem for multi-cryptographic job collaborative service was modeled as a multi-objective optimization workflow scheduling problem.A two-stage“select-sort”scheduling algorithm was proposed.In the selection stage,the improved NSGA-Ⅲ algorithm was employed to select computing units for cryptographic computing jobs,and in the sorting stage,the execution order was determined based on the urgency of jobs.Simulation results demonstrate that the proposed algorithm outperforms traditional scheduling algorithms in terms of energy consumption,migration costs,and adaptability to transient surges in loads.

关 键 词：云计算 密码按需服务 作业流调度 NSGA-Ⅲ 

分 类 号：TN92[电子电信—通信与信息系统]