论文部分内容阅读
针对动态网格资源服务的不确定性问题,提出一种可量化分析资源服务QoS(quality of service)的多资源协同预留策略.该策略基于对运行在资源上的网格任务QoS指标分析,得出QoS满意度量化、归一化方法,建立资源服务QoS与预留容量之间的函数关系,并以市场经济环境为背景,分析任务费用约束下资源价格与预留容量之间的关系,求解得出可均衡负载的多资源节点协同预留方案.理论分析给出了策略的有效性证明和算法,仿真实验采用真实网格系统中的任务负载信息作为实验负载,在较大规模的模拟网格系统中检验了所提出的预留策略的性能表现.实验结果显示,该策略在接纳任务数、资源利用率和任务违约率方面的性能表现显著优于传统的预留策略.
Aiming at the problem of the uncertainty of dynamic grid resource service, this paper proposes a multi-resource collaborative reservation strategy that can quantify and analyze the quality of service (QoS). This strategy is based on the analysis of QoS indicators of grid tasks running on resources, The quantitative relationship between resource price and reservation capacity is analyzed under the background of market economy environment. The relationship between resource price and reservation capacity is analyzed under the condition of market economy. The multi-resource node collaborative reservation scheme is obtained, which can be load balanced.A theoretical analysis is given to prove the effectiveness of the algorithm and the algorithm. The simulation experiment uses the task load information in the real grid system as the experimental load, The performance of the proposed reservation strategy is tested in the grid system.Experimental results show that this strategy outperforms the traditional reservation strategy in terms of the number of tasks accepted, the resource utilization rate and the default rate of the task.