网格计算中基于QoS的资源调度优化模型

QoS Based Resource Scheduling Optimization Policy in Computational Grid

文中对网格计算的多维度QoS进行量化,建立基于多维度QoS的网格资源调度优化模型.给出QoS维度的定义,这些维度包括花费、完成时间、信用度等.采用离散效用函数来描述用户的每一维QoS需求,每一维QoS都有自己的组合权重,网格服务的组合QoS即为综合效用函数.通过对网格资源约束和QoS服务需求进行数学抽象,将多维度QoS的资源调度的概念模型转化为一个多约束优化模型,该模型能够反映用户在服务质量上的不同要求.

This paper presents a QoS based grid resource scheduling optimization model by measuring multiple dimension QoS in grid computing. The definition of QoS dimensions is given in different aspects as cost, completion time and reliability. The paper applies discrete utility functions to describe each user with its own dimensional QoS which consists of composition weight and grid service＇s composite. QoS can be represented by collective utility function. Through mathematically abstracting grid resource constraints and QoS requirement, the multiple QoS based resource scheduling can be formulated as a multiple constraints optimization model which reflects grid user＇s diverse QoS requirements.