Study Of Network Traffic Management Based On Netflow

With the ever increasing size and complexity of network, network flow visualization is an indispensible functionality of network management software. Nowadays most net-flow visualization management software only possesses the net-flow statistics diagram functionality while lacking the ability to analysis the net-flow distribution combine with network topology. In order to solve this problem, in this paper we designed and implemented a prototype system which can analysis the net-flow distribution while taking into account of the network topology. This system can swiftly identify elephant flow, and show it on the network topology view.In order to meet the demand of the prototype system, we design a net-flow-oriented get elephant flow algorithm. This algorithm identifies net-flow by three iterations of filtrating. We demonstrated that this algorithm can identify the net-flow elephant flow in a smart way.We also developed a subnet-base host computer network topology discovery technique. It can get the network topology via applying different network topology techniques. After that we designed a host-to-host path deduction method, we showed the method is efficient via experiments.In the last part of the paper, we give a brief description of the design and the data interface of the system. By analysis the network flow and topology in a synthesis way, the system can identify different types of net-flow and visualize it on the network topology.We showed that this system is useful to analysis net-flow.