A Filtering Method Research Of Customizable URL Classification

Over the past decade, with the rapid development of the Internet, the network security issues are also increasingly serious. The main means of current Internet firewall is content-based filtering and URL filtering. Content-based filtering is still unsatisfactory in the processing speed, and URL filtering is very popular filtering methods in the vast majority of the corporate firewall. In the face of the huge Internet information resources, classification and selective filter of Internet content is becoming increasingly important. The filter services of customizable URL classification are popular in the current network security needs.According to the different filter demonds of network security among different users, the thesis, basing on the trend of development of next-generation Internet firewall, proposed a new filter method. Facing to customizable classification of URL, the method can be applied generally to the appliances of network security. The filter method firstly import URL classification information corresponding to the size of the storage capacity of network security appliance, which contains most popular and frequently accessed URL. And then designs the method of selection rule, the way of information organization as well as the reliable consideration of classification information. Based on the pre-definition classification library, the method provides customizable URL rule, classification, filter action and strategy, it also supports all kinds of customizable functions such as quotation to pre-definition URL rule and classification, addition of new URL rules classification and so on. The network manager can make the strategy about data flow control flexibly according to the need of development. The Storage of network equipment itself is limited. With regard to unknown URL, It gains classification information from remote sever with large storage in the way of cloud query. According to the custom policy of user, the method can face on the sountless URL flow, complete flexible, effective and stable data flow monitoring and realize reliable filter which is user-oriented and custom safe. The thesis has also designed the inquiry process which is safe and fast, and established the dynamic learning mechanism of classification to avoid much of unnecessary inquiry.Aimming at the method above, the thesis designed filter system, carried on the engineering design and the code realization. After real data flow control test deployed on network equipment, it shows that the filter method is feasible and have high value of practical application to future internet security.