摘要
针对云系统规模庞大、构成复杂、动态性突出、层次关联性强而难于建模评估的问题,提出一种基于排队Petri网的云系统评估模型QPNC;QPNC结合了排队论和Petri网理论特点,模型具备较强的定量评价和行为描述能力,能够很好地对复杂云系统进行有效建模和模拟;基于上述模型,进一步提出并完善云系统的定量分析、评估体系,仿真并模拟了大规模并行环境下云系统的动态服务效果;实验结果表明,QPNC能够有效反映出各种云系统架构在性能和服务等方面特征,对云系统的各种动态服务行为具有很高的仿真度,为设计构建更加高效、更具针对性的云系统提供了定量分析支持和理论依据。
To solve the lack of modeling and evaluation solution in cloud system with large scale, complex structure, significant dyna- mism and Strong correlation among hierarchies, we propose a novel Queuing Petri Net based Cloud system evaluation model （QPNC）. QPNC combines Queuing Theory and Petri Nets and aims at taking their advantages. The model is adaptive to the quantitative evaluation and activi ty description, thus can effectively model and simulate the complex cloud system. Based on the model, we further propose and improve the system of quantitative analysis and evaluation for cloud system and simulate the dynamic service effect of cloud system under massive parallelism cases. The experimental results show that QPNC can effectively reflect the characteristics of cloud system in performance and service and highly simulate the varied dynamic service behaviors of cloud system, which provides quantitative analysis and theoretical basis for designing and building more effective and targeted cloud system.
出处
《计算机测量与控制》
2015年第8期2878-2881,共4页
Computer Measurement &Control
基金
国家自然科学基金(U1304603)
关键词
云系统
排队论
评估模型
digital watermark
wavelet packet
adaptive
feature of texture
