Abstract: Probe Gap Model(PGM) is a lightweight and fast available bandwidth estimation method, but the algorithms based on PGM are lack of systematic and in-depth assessment. This paper analyzes the main factors effecting the accuracy of PGM’s algorithms, proposes a new classification method of the end-to-end path available bandwidth, selects representative Spruce and IGI and evaluates the impact of the vital factors including cross traffic on the accuracy of algorithms systematically through theoretical derivation and simulation test. The research provides a reference to improve the design of algorithms and the accuracy of algorithms.

Key words: available bandwidth, Probe Gap Model(PGM), evaluation

摘要： 探测间隔模型(PGM)是负载小、速度快的可用带宽度量方法，但目前针对PGM的算法还缺乏系统、深入的评估。针对上述问题分析影响PGM算法精度的主要因素，提出端对端路径可用带宽的新型分类方法，选取PGM模型中较典型的Spruce和IGI算法，通过理论推导和仿真测试，评估不同背景流量类型等因素对算法精度的影响，为改进PGM模型的算法设计、提高算法精度提供参考。

关键词: 可用带宽, 探测间隔模型, 评估

CLC Number: 

• TP393