基于排队模型的包链测量方法特征分析  

Analysis on characteristics of packet train measurement methods based on queuing model

作　　者：何晓桃[1] 黎文伟[2] 

机构地区：[1]广东工业大学计算机学院,广州510006 [2]湖南大学信息科学与工程学院,长沙410082

出　　处：《计算机应用》2011年第7期1920-1923,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(60703097)

摘　　要：基于单跳路由器的排队模型,研究了受背景流量影响下的探测包链排队行为。随着背景流量影响程度的不同,分析得出无背景流量排队、短背景流量排队、长背景流量排队和多背景流量排队等四种具有不同时延变化分布特征的典型探测包链排队行为。仿真实验通过调整包链发送速率,模拟了不同的背景流量影响程度,发现探测包链的包时延变化有不同的分布特征,与分析的四种典型探测包链排队行为特征基本一致。Understanding the queuing characteristics of probing packet trains is the basis for designing bandwidth measurement methods.Based on a single hop queuing model of routers,the queuing characteristics of probing packet trains under the impact of cross traffic were studied.Four typical queuing behaviors of packet trains which have different packet delay variance distribution,i.e.no cross traffic queuing,short cross traffic queuing,long cross traffic queuing,and multiple cross traffic queuing,were analyzed,with different degrees of cross traffic load.The experiments simulated different degrees of cross traffic load,by setting different probing packet train sending rates,and found that the packet delay variance of probing packet trains has different distribution characteristics,which is consistent with the analytical results.

关 键 词：包链测量 带宽测量 背景流量 排队分析 

分 类 号：TP393.07[自动化与计算机技术—计算机应用技术]