The Research Of Web Traffic Hijacking Discovery Technology

With the rapid development of the Internet and the wide application of information technology,the network has penetrated into all walks of life and people's daily life.In recent years,network attacks have emerged in endlessly.Web traffic hijacking is one of the most common types.If web traffic is hijacked,it will have a great impact on service providers,browsers and even the whole Internet.But at present,there are few research on Web traffic hijacking detection technology,and the false alarm rate is high.Therefore,it is of great significance to conduct in-depth research on Web traffic hijacking detection technology.Web page similarity comparison algorithm can draw the conclusion whether the web page has been tampered with.The comparison algorithm needs to find the history information of the web page according to the URL,and then compare the history information with the current information.Therefore,the URL memory index plays a vital role in the process of web page information storage and search.In this paper,we will make a thorough study on the optimization of web page similarity comparison algorithm and URL memory index.Firstly,we optimize the similarity comparison algorithm of web pages.In this paper,HTML is studied in depth,and the concepts of HTML element's self-defined classification and influence factor are put forward to assign weight to different levels of DOM tree,and the structural similarity is obtained.For web content,the concepts of text set and the degree of variation of text set are put forward,and the content similarity is obtained.Combining with the features of website updating and tampering,this paper puts forward some concepts such as important attributes,and finally obtains the similarity comparison algorithm,and proposes a judgment model of Web tampering based on multi-point similarity comparison.In the process of implementation,virtual DOM is used instead of too heavy DOM.By comparing the experimental data,the advantages of the similarity comparison algorithm proposed in this paper are proved to be low false alarm rate.Secondly,optimize the URL memory index.On the basis of TRIE tree,this paper uses the characteristics of URL segmentation to hierarchically process the index structure;combines multiple binary nodes as set nodes,and extracts them in parallel using SIMD technology;provides a variety of storage structures,and set nodes can choose the most suitable storage structure;each set node is fan-out fixed,maintaining a small and stable tree height.By comparing the experimental data,it is proved that URLS-DNT structure consumes less space,performs well in search and insertion,and has low and stable tree height,and can achieve efficient URL index.