【作者】 谷上宇； 赵国锋； 唐红；

【Author】 Gu Shang-yu, Zhao Guo-feng, Tang Hong(Chongqing University of Posts & Telecommunications, Chongqing 400065. China)

【机构】 重庆邮电学院；

【摘要】 对于高速IP骨干网,传统的基于SNMP/RMON机制进行网络内部链路流量统计测量的方法,会在网络核心路由/交换节点设备处产生较大的处理负担,降低处理器性能,从而影响数据包的快速转发.本文提出了一种利用边界测量估计网络内部链路流量的模型,利用抽样方法,在网络边界接入链路处进行测量,得到源目的SD(Source/destination)节点对间的流量,然后结合路由信息,计算得出网络内部链路的流量.该模型的优点是仅在网络边界处进行测量,而不涉及网络桉心,因此引起的网络开销较小.性能仿真结果显示,测量模型是有效的.更多还原

【Abstract】 The traditional approach to network internal links traffic using SNMP/RMON can neither decrease processing load of the kernel nodes equipments in high speed IP backbone network nor transfer complex tasks to network edges. Accordingly, we propose a model of estimating network links traffic through edge measurement By means of sampling, in the access-links on the verge of network, we can get the traffic between the SD nodes. Combining it with trajectory information, we calculate the traffic of internal links in the network. The advantages of the model are that, not involved in core network, we only need to measure traffic on the boundaries of the network. So its implementation cost and error are small. The emulating result shows this measurement model is feasible.更多还原