计算机科学与探索 ›› 2012, Vol. 6 ›› Issue (6): 531-544.DOI: 10.3778/j.issn.1673-9418.2012.06.006

• 学术研究 • 上一篇    下一篇

运用云模型评价复合网络情境的方法研究

陈宁江1+,张莉莉1,张文博2   

  1. 1. 广西大学 计算机与电子信息学院,南宁 530004
    2. 中国科学院 软件研究所 软件工程技术研究开发中心,北京 100190
  • 出版日期:2012-06-01 发布日期:2012-06-01

Applying Cloud Model on Assessment Scheme of Complex Network Context

CHEN Ningjiang1+, ZHANG Lili1, ZHANG Wenbo2   

  1. 1. School of Computer, Electronics and Information, Guangxi University, Nanning 530004, China
    2. Technology Center of Software Engineering, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China
  • Online:2012-06-01 Published:2012-06-01

摘要: 由于网络的开放性、动态性、难控性等特征,网络情境参数的取值及其变化表现出模糊性、随机性等不确定特征,对Internet中软件系统的运行和管理具有重要的影响。在对实际网络应用需求分析的基础上,利用云模型在定性概念与定量数值之间转换的优点,研究了一种基于云模型的复合网络情境评价机制。它对多维网络情境因素的特征进行了刻画,能够实现多维情境参数与运行时网络环境评估之间的量化映射,较好地处理情境参数的不确定性和模糊性;通过加入用户误差承受度,动态调整样本数量,使建模结果更为稳定和准确;实现了在参数级和群体级两个粒度的网络情境参数特征建模,较好地表征了网络情境的整体评价,有助于避免配置抖动问题。该方法的效果通过系列仿真实验得到了验证。

关键词: 网络情境, 云模型, 情境建模

Abstract: As Internet is open, dynamic and difficult to control, the values of network contextual parameters and their changes show uncertain characteristics, which impacts importantly the execution and management of software systems on Internet. Based on the analysis of practical Internet application requirements, this paper proposes a new complex network context assessment scheme, by making use of the advantages of cloud model in transformation between qualitative concepts and quantitative values. It profiles the features of multi-dimensional network context factors, and achieves the quantitative mapping between the multi-dimensional contextual parameters and the assessment of run-time network environment. By adding the user tolerant error into calculation to adjust the number of samples dynamically, more stable and accurate assessment results can be obtained. The scheme achieves the modeling results at both parameter-level and group-level, which brings a better overall evaluation of the network status as well as helps to improve the correctness and efficiency of dynamical software reconfiguration. The effect of the scheme has been verified through a series of simulation experiments.

Key words: network context, cloud model, context modeling