Real-time volume control for interactive network traffic replay

  • Authors:

  • Weibo Chu;Xiaohong Guan;Zhongmin Cai;Lixin Gao

  • Affiliations:

  • MOE KLINNS Lab, Xi'an Jiaotong University, Xi'an, China;MOE KLINNS Lab, Xi'an Jiaotong University, Xi'an, China and Center for Intelligent and Networked System and NLIST Lab, Tsinghua University, Beijing, China;MOE KLINNS Lab, Xi'an Jiaotong University, Xi'an, China;Department of Electrical and Computer Engineering, University of Massachusetts, Amherst, USA

  • Venue:
  • Computer Networks: The International Journal of Computer and Telecommunications Networking
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Interactive network traffic replay is a new traffic generation tool very useful for testing in-line networking devices. It provides both a realism of traffic contents and a realism of traffic behaviors. But the lack of effective mechanisms to control important features of the generated traffic severely limits its capability of performing accurate, systematic and in-depth testing. This paper aims to improve controllability in interactive network traffic replay by studying the problem of controlling volume of the generated traffic. Due to complex traffic generation mechanism, controlling traffic volume in interactive network traffic replay is an interesting and challenging new problem. In this paper, we formulate the volume control task as a dynamic target tracking problem and analyze the influencing factors and basic control properties of the system. We propose two approaches, from both model-based and model-free perspectives, for solving the problem. The model-based approach relies on a probabilistic model that explores packets processing mechanism and converts the problem to a state prediction problem, while the adaptive optimal control approach uses process input/output information to regulate the system and converts the problem to a parameter estimation problem. Both of the two methods are validated via extensive experimental studies with real-world traces. Results indicate that our methods can track target output traffic volume effectively under a wide range of network conditions.