Tomography技术中的多参数网络拓扑推断

为了减少拓扑推断中采用单一参数在推断时的限制，提出了一种结合端到端时延和丢包率的算法，该算法在应用时仅需适量的计算而无需增加测量流量，并且该算法能适应不同负载的网络环境。理论分析了多参数拓扑推断算法的正确性并通过网络仿真器(NS2)进行了仿真，仿真结果表明多参数的网络拓扑能在不同负载的网络环境下准确的推断网络拓扑。

Multiple Parameters Network Topology Inference Based on Tomography

In order to reduce the limits in topology inference of one parameter, a multiple parameters network topology inference method, combining end to end delay and loss, was proposed. When applying the multiple parameters based network topology inference, no additional traffic was needed; the inference method can be applied in complex networks of different loads. The multiple parameters based inference method was analyzed theoretically and validated through simulations by network simulator, version 2 (NS2). The results of simulation illustrated that the multiple parameters based inference method could infer network topology correctly when faced with different loaded networks.