Situation Assessment Of Network And Visualization In AD HOC Network

Situational awareness is one of the key technologies for network management. This technique makes it possible to mine deep information, prospect the trend of states in the network. As one of the key technologies of Situational awareness, situation assessment can analyze current statuses and make us able to prepare for network congestion and security events. Thus it can ensure normal operation of the network. Based on the previous studies of situation assessment, this dissertation introduces some analysis techniques, like complex network and neural network, to develop situation assessment algorithms further. Meanwhile, the paper goes on to study the application of situation assessment, using visualization to construct rich clients, which achieves a network situation assessment software block and assesses network states in real web environment.Firstly, introducing the concept of betweenness in the area of networks, we propose a situation assessment algorithm, which is used to assess the importance of network nodes and links, according to the concept of betweenness and current network characters. Through analysis of the network topology, new indicators are given in terms of network construction, namely the importance of nodes and links based on topology betweenness. On the other hand, we introduce the concept of betweenness to analysis of network flow data, according to the way betweenness is used for importance assessment, and propose dynamic indicators of network importance which fit for web dynamic monitoring features, namely the importance of nodes and links based on network flow betweenness. The paper states the relevant theory and gives algorithm detail, supporting the application based on the analysis of the simulation.Secondly, the paper integrates neural network algorithm, applying the BP neural network to assess network importance. Developing the original algorithm, a new algorithm is propsed to help analyse networks. Through training of neural network, the algorithm can assess networks efficiently, prospect network states and offer analysis data in time. The paper interprets relevant theory of this part, giving idea of the algorithm design detail, after which the reasonableness has been proved via simulating a variety of indicators.Thirdly, visualization has been used in assessment of networks. Combining rich client presentation techniques, such as ExtJS and Geoserver, we constitute a software block of situation assessment, realizing view presentation of network state data and the corresponding data analysis. With visualization techniques, it’s easier to understand the information offered by software and views of network analysis is increased, which brings administrators further insight into operation states of networks. This practice also increases the degree of visualization in network management. At the same time, we adjust the formation of data in combination of application requests and go further in improving usability of visualization in the field of situational awareness.