Research On Method Of Feature Matching For HTTP Protocol

With the rapid development of the Internet and mobile Internet,network traffic is gradually becoming more and more complicated and massive.In the face of today's network traffic,the traditional matching method has gradually becoming unable to adapt,due to its flexibility in the matching problem,which is mainly reflected in two points:1)There are usually some relative position relationship in the network traffic between keywords.This kind of relative position relation often embodies some kind of logical meaning in it,the traditional string matching algorithm can't meet this demand;2)With the development of the network,the security rules in a specific traffic area show an explosive growth trend and traditional string matching algorithm is often universal,which means that in the face of a specific traffic,it can't make targeted optimization to adapt to this large-scale string matching in specific areas.From the above two perspectives,this article proposed HTTP multi-domain combination of expression matching and large-scale URL matching string problems,combined with HTTP protocol,which is the most widely used protocol in the Internet.Specifically including the following aspects:1)Analyze the characteristics of HTTP header information and the matching of HTTP header information,we propose multi-domain combination expression matching problem,and divide this expression into "logical and" expression and "partial and" expression.2)By referring to the semantic logic of the regular expression itself,the multi-domain combined expression is transformed into a matching problem of a special regular expression containing only".*".At the same time,drawing on the application of regular expression in the field of network security,in order to obtain the tradeoff between the matching speed and the memory consumption,we adopt the overall architecture design of hybrid automaton.And grouping hybrid automata and automata's alphabet reduction optimization methods were applied.On this basis,we given the overall matching framework of this algorithm.3)Aiming at the problem of large-scale URL matching,based on the existing pattern matching algorithm and combined with the characteristics of URL data,a variety of optimization methods are given to reduce the probability of hash conflicts and reduce the number of exact checks during collision.All these methods have improved the overall matching speed.