引入多QoS开销适应度的云计算任务权衡调度  

Cloud Computing Task Trade-off Scheduling Introduced in QoS Overhead Fitness Computing

作　　者：王捷[1] 王顺平[2] 

机构地区：[1]郑州轻工业学院计算机与通信工程学院,郑州450002 [2]河南科技学院,河南新乡453003

出　　处：《科技通报》2015年第6期154-156,共3页Bulletin of Science and Technology

摘　　要：提出一种引入QoS开销适应度运算的云计算任务权衡调度算法,首先进行了支持多QoS因素任务调度问题描述与网格拓扑结构构建,进行云计算任务权衡调度对多用户QoS偏好的影响力数学度量,通过多QoS开销适应度运算的引入,根据计算资源的成本和数据传输时间,来确定分配任务的位置。为了适应云存储中的多QoS偏好,重新定义PSO的适应度函数,实现任务权衡调度算法的改进。通过仿真实验研究得出,采用该算法对云计算任务节点的聚类准确性较高,进行任务调度中的实时性好。通过多QoS偏好分析,引入QoS开销适应度运算,用户满意率有明显上升,适应度函数随不同类别任务变化,有效地反映不同类型任务的QoS偏好。展示了较好的云计算任务权衡调度性能。A new QoS overhead fitness computing cloud computing task scheduling algorithm first constructs a trade-off is proposed, support for multiple QoS factors of task scheduling problem description and mesh topology structure of cloud com?puting, task scheduling metrics tradeoff influence of mathematics to multiuser QoS preference, adaptation through multiple QoS overhead of the introduction of operations, according to the cost of computing resources and the data transfer time, to determine the allocation of the location for the task. In order to adapt to the cloud storage of multi QoS preference, redefin?ing the fitness of PSO, improved task scheduling algorithm to achieve balance. Through the experimental study of simula?tion, the algorithm can use this algorithm in the face of cloud clustering accuracy higher calculation task node, for real-time task scheduling of a good. Through the multi QoS preference analysis, introducing the QoS overhead fitness operations, cus?tomer satisfaction rate increased significantly, the fitness function with the change of different types of tasks, it can effective?ly reflect the preferences of different types of task QoS, it show good cloud computing task scheduling performance.

关 键 词：QoS开销 满意度 权衡 云计算 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]