包对探测统计模型与测量精度分析

Statistical modeling and accuracy analysis of available bandwidth measurement

包对模型(PGM)是探测可用带宽的主要方法。该方法存在以下问题,影响其推广应用:模型没有精确的数学模型描述,原有模型假设流量包长看做无限小,与实际网络情况不符合;没有定量方法计算探测精度、样本容量,给测量带来不便;没有分析互联网数据包突发性对测量带来的影响。为解决以上问题,通过对包对模型进行数学建模的方法,证明了模型正确性;定量分析了测量精度;确定了样本数计算公式;分析了数据包突发性对探测造成的影响;最后对以上分析进行了仿真实验验证,提高了可用带宽测试技术的精度和可信度。

Available bandwidth measurement is widely applied in a dependable network. This paper proposed probe gap model （PGM） as a lightweight and fast available bandwidth estimation method. But a big problem of PGM was its poor accuracy in certain network conditions. The accuracy could＇ t be trusted. An important assumption behind the PGM was that the model should be statistical robust even the cross traffic follows do NOT follows fluid model. But no verification could prove this as- sumption. These problems made tools based on PGM could＇ t be widely used. Constructed a statistical model for PGM. Based on this insight, presented mathematical verification and accuracy analysis for PGM. And also use modeling, measurements to characterize the interaction between probing packets and bursting network traffic. NS simulation shows the modeling and the ac- curacy analysis is successful. By the model, can know how much can trust the probing results.