基于任务概率聚簇的工作流云费用优化调度算法  

A Cloud Workflow Scheduling Optimization Algorithm Based Task Probability Clustering

作　　者：王彬 Wang Bin(Chongqing College of Electronic Engineering,Chongqing 401331,China)

机构地区：[1]重庆电子工程职业学院

出　　处：《信息与电脑》2019年第22期17-19,共3页Information & Computer

摘　　要：由于云资源具有弹性伸缩、按使用量计费等优势,科研机构将工作流部署在云环境执行。工作流任务在云计算环境执行过程中会产生不可避免的系统开销。针对实际运行环境系统开销过大的问题,笔者提出基于聚簇概率的任务概率聚簇算法,平衡每层任务的资源需求,提高资源分配粒度,降低系统开销和降低公有云执行费用。实验表明,该算法能够有效降低公有云花费。Because of the advantages of flexible scaling of cloud resources and charging according to usage,scientific research institutes deploy workflow in the cloud environment for execution.Tasks will inevitably incur system overhead during the execution of cloud computing environment.Aiming at the problem of excessive system overhead,a task probability clustering algorithm based on clustering probability is proposed to solve system overheads in real system,so that horizontal task clustering algorithm not only balance resource demands,but also reduce system overhead and public cloud cost.Simulation results prove that TCBP can reduce cost.

关 键 词：云工作流 任务概率聚簇 系统开销 费用优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]