Network Security Situation Awareness And Trend Analysis System

With the advent of the information age, computer and communications technology has developed rapidly, and they have an increasingly important impact on people's lives. At the same time, computer networks and operating systems'own loopholes exposed more and more to the people, endless criminal activities happen by using computer networks, not only caused enormous loss of property, or even a threat to the country's overall information security. Therefore, in order to protect the network security services, the problems of timely understanding of the current state of the network and forecasting its development trend is urgently to solve. In order to predict the trend of the network, with the network's own characteristics, the paper put forward four index systems that are interrelated, and cover of the entire network at all levels. These indexes quantitatily descrip the various parts of the characteristics of the network and through comprehensive compute find out the network security value, which is used to descrip the network's overall state. After we get the set of the network security values, we can base on the analysis of time series data sets to forecast the network state.This paper includes the following works: (1) Post a network security value prediction algorithm, which is based on the RBF neural network. (2) Design and implement a network security situational awareness and trend analysis system.Against the network trend forecast, this paper mainly based on the RBF neural network security value prediction methods. Neural network is a technology system, which simulated human's brain structure and function. It is a large-scale parallel nonlinear dynamic system, which uses a large number of nonlinear parallel processors to simulate people's neurons, and also uses flexible intricate inter-processor connections to simulate the human's brain neurons synapse. Compared with the traditional forecasting methods, neural network has highly nonlinear computing and imaging capability to function with any degree of accuracy approximation relations, and has a strong ability to forecast more accurate. On the basis of analysis traditional neural network algorithm, we find out its inadequate and make some improvements, and make use of it in the network situation prediction. Subsequently, we conducted on the system design, including the overall structure design, database design and interface design. And focus on the achievement of the network security value forecast subsystem. The subsystems include cache, data preprocessing, RBF neural network structure identification, and the right value adjustment four modules. Cache module is mainly responsible for the optimization of data storage, data preprocessing module is responsible for the preparation of sample data, the latter two modules are responsible for analyze the history data in order to obtain the predictive value. Finally, we tested the system. In the actual network environment through the operation, we found that the prototype system can effectively predict the network situation and to specify its development trend.