电信科学 ›› 2011, Vol. 27 ›› Issue (9): 66-72.doi: 10.3969/j.issn.1000-0801.2011.9.017

• 研究与开发 • 上一篇    下一篇

大规模网络模拟的背景流量建模

钱亚冠1,王滨2,关晓惠3   

  1. 1 浙江科技学院理学院 杭州310023
    2 浙江大学计算机学院 杭州310027
    3 浙江水利水电专科学校计算机系 杭州310018
  • 出版日期:2011-09-15 发布日期:2017-06-22
  • 基金资助:
    国家科技支撑计划基金资助项目;国家“973”计划基金资助项目;浙江省教育厅科研项目基金资助项目

Background Traffics Modeling Applied in Large-Scale Network Simulation

Yaguan Qian1,Bin Wang2,Xiaohui Guan3   

  1. 1 Science College,Zhejiang University of Science and Technology,Hangzhou 310023,China
    2 The College of Computer Science,Zhejiang University,Hangzhou 310027,China
    3 Department of Computer Science,Zhejiang Water Conservancy and Hydropower College,Hangzhou 310018,China
  • Online:2011-09-15 Published:2017-06-22

摘要:

大规模网络模拟的一个巨大挑战就是对计算资源的巨大需求,因此建立适合大规模网络模拟的背景流量模型具有十分重要的意义。本文提出的一种背景流量模型可实现3个目标:真实性、高效性和对网络动态的反馈性。该模型利用ON/OFF模型表达两种典型的应用(Web和P2P)的用户行为模型;在TCP层上采用fluid的抽象流量表示,以提高模拟效率;用微分方程表达拥塞避免行为,使得流量源具有对网络的响应能力。最后通过DaSSFNet建立了模型原型,通过实验模拟,该模型框架可以很好地表达自相似的流量特征,同时在模拟效率上也远远高于包级流量模型。因此,该模型适合作为大规模的网络模拟环境的背景流量发生器。

关键词: 大规模网络模拟, 背景流量, ON/OFF模型, 流体流

Abstract:

As the large requirements of computing resources for large-scale simulation are becoming a challenge,it is needed to construct a proper background traffic model to overcome these difficulties. We proposed a background traffic model frame to realize the following objects:realistic,efficient and responsive. ON/OFF models are used to describe the user behaviors by two typical applications(Web and P2P). Traffic is abstracted as fluid to get more efficient than packet level presentation and the ODE(ordinary differential equation)is used to depict the congestion avoidance functionality. Finally,a prototype model was constructed and some experiments are conducted. The results imply that the aggregated traffic from the model presents expected self-similarity and the efficiency is more obvious than packet level traffic simulation.

Key words: large-scale network simulation, background traffic, ON/OFF model, fluid flow

No Suggested Reading articles found!