Research On Network Topology Optimization Detection And Recognition Method

With the development of network technology and the popularity of network services,the Internet has become an indispensable infrastructure in social life.However,the everincreasing demand for network makes the growing network size and complexity of the structure.As the ”skeleton” of the network,the network topology has important significance for the construction,management,and security of the network.In order to obtain the network topology,a great deal of researches have been done at home and abroad.There are a lot of effective and novel research results in the network detection methods and network topology identification.However,few researches have been done on the deployment of probe nodes,and there are some outstanding difficulties in network topology identification,which leads to unsatisfactory results of network topology detection and restoration.In a complex network environment,the deployment of probe nodes is an important factor that affects the validity of the data and the cost of detection.The method of network topology identification is a key factor that transforms the original data into the network topology.Therefore,the research work of this thesis will focus on these two aspects and propose the following three related methods:1.Proposed topology detection node deployment method based on topological feature constraints.For actual deployment of topology detection nodes with unknown global topology and known partial topology,this thesis first uses a partial topology to generate the similar global topology,and then deploys nodes in the global topology,and finally selects the final deployment results based on deployment plans of different topologies.In order to achieve the above-mentioned thinking,this thesis first analyzes the similarity features between partial and global network topology,and proposes a global topology generation model based on hierarchical,power-law and self-similarity.Second,it also analyzes the detection characteristics of the nodes under Traceroute,and proposes a global topology deployment method based on mixed weights.Compared with other deployment methods,the proposed method can significantly reduce the number of probe nodes and the cost of network detection,and reduce the performance impact on the measured network.2.Proposed anonymous router identification method based on multi-source information fusion.Because the network tomography is not affected by the cooperation of intermediate nodes,this thesis combines the network tomographic results with Traceroute information to identify the anonymous IP in Traceroute.Firstly,the mapping relations between the two detection methods are analyzed.Secondly,the mapping relationship between the two detection methods is transformed into a mathematical model of integer programming by finding constraints.Finally,the anonymous IP in Traceroute is merged through the results of the mathematical solution.This method fuses the data information of the two detection methods,inherits the advantages of the two,and overcomes their respective shortcomings.This method has a high accuracy rate on the final result,so it resolves the identity problem of the anonymous router to some extent.3.Proposed router alias resolution method based on multi-source information fusion.Firstly,the traditional alias resolution algorithm APAR is analyzed and we found that its data information is too single and it relies on the correct IP subnet information.To overcome these shortcomings,the topology information in the route message is added as a data source,and the rules for extracting IP addresses,IP subnets,and IP aliases in the packets are summarized for three routing protocols: OSPF,RIP,and EIGRP.Finally,these information is integrated into each step of the APAR algorithm to enhance its alias resolution.Compared with the traditional method,this method overcomes the existing disadvantages and greatly improves the accuracy of alias resolution.In order to verify the effectiveness of algorithms,this thesis conducts comparative experiments or simulation experiments on these algorithms at the same time.For the node deployment algorithm proposed in this paper,we use a large number of random topologies as the test environment to compare with other node deployment methods in the number of deployment nodes,detection range,and detection cost.For the proposed anonymous router identification algorithm and alias resolution algorithm,we use NS-3network simulator to simulate the network topology detection and identification of these two methods,and study the factors that may affect the experimental results.The above experimental results show that the proposed algorithms in this paper are effective and can solve some problems in the network topology detection to a certain extent.