Open Access Article

Title: End-to-end available bandwidth estimation using HybChirp

Authors: Wenzhen Chi; Tao Zheng; Yi Xie; Zhongwen Li; Yijiang Chen

Addresses: School of Information Science and Engineering, Xiamen University, Xiamen, 361005, China; ShenZhen Research Institute of Xiamen University, ShenZhen, 518000, China ' School of Information Science and Engineering, Xiamen University, Xiamen, 361005, China; ShenZhen Research Institute of Xiamen University, ShenZhen, 518000, China ' School of Information Science and Engineering, Xiamen University, Xiamen, 361005, China; Fujian Key Laboratory of Sensing and Computing for Smart City, Xiamen University, Xiamen, 361005, China ' The Computer College, Chengdu University, Chengdu, 610106, China ' School of Information Science and Engineering, Xiamen University, Xiamen, 361005, China

Abstract: The available bandwidth is a crucial metric of networks performance which is applied in congestion control, route selection, traffic analysis and QoS management. PathChirp is a well-known tool to estimate available bandwidth based on an approach of self-induced congestion, which exponentially increases the rate of probing packets in each packet train. But the inappropriate gaps of probing rates often result in a large deviation of estimated available bandwidth. In order to solve this problem, we propose HybChirp, an active probing tool which uses linearly spaced probing packets (lin-chirps) and exponentially spaced probing packets (exp-chirps) in combination. HybChirp improves PathChirp by adopting the hybrid of lin-chirps and exp-chirps, while inherits the advantages of PathChirp such as the short convergence time and low overhead. The NS-2 simulation results have shown that HybChirp outperforms PathChirp and Pathload in terms of accuracy, overhead, convergence time and adaptivity.

Keywords: available bandwidth; bandwidth estimation; PathChirp; linear probing; exponential probing; HybChirp; network performance; congestion control; convergence time; overheads; simulation; accuracy; adaptivity.

DOI: 10.1504/IJCSE.2016.076954

International Journal of Computational Science and Engineering, 2016 Vol.12 No.4, pp.360 - 369

Received: 05 Apr 2015
Accepted: 22 Aug 2015

Published online: 08 Jun 2016 *