The Research On Comprehensive Evaluation Method For Network Security

Network security assessment is not a new concept. It can date back to the year 1985, when the United States Department of Defense issued the first evaluation criteria–Trusted Computer System Evaluation Criteria. With more than 20 years of research and development, a lot of security evaluation criteria and evaluation method have been generated so far. And, most of these existing evaluation methods focus on one particular aspect of network security, such as technology, management, process, person and so on. However, network security is influenced by many complicated factors, combine all these factors to evaluate will get a more scientific and reasonable result.Fuzzy Comprehensive Evaluation absorbs the advantages of fuzzy set theory in dealing with the subjective factors, solves the fuzziness of network security factors, besides, it can make comprehensive analysis of various factors to make an overall evaluation. For these reasons, fuzzy comprehensive evaluation can effectively solve the subjective and one-side problem that exist in the traditional assess methods.The selection of factors and determination of factor weight is crucial to the fuzzy comprehensive evaluation method. This paper established a three-level network security evaluation factor system framework, adopt the modified Delphi method to select the network security factors, in determination of factor weights, to against the shortcomings of the traditional AHP weight determination method, this paper introduce a improved method based on triangular fuzzy numbers, named IFAHP, which can solve the problem of subjective and consistency test in AHP method effectively. Finally, a network security model is established, which utilizes the previous evaluation factor selection and factor weights determination method, using 2-level fuzzy comprehensive evaluation method to assess the security of the network system. In addition, a network security evaluation prototype system based on the model has been developed to verify the feasibility of the model through experiments.