The Research Of Pattern Matching Algorithms Based On Regular Expression Rules Set

With the use of regular expressions in the network security system and the application of a variety of services is more and more wide, the system uses a core which consists of regular expression matching algorithm, to detect the attack characteristic in the packet payload. In recent years, a lot of studies focused on the large-scale regular expression rules set, how to effectively reduce the DFA storage overhead. In the modern network intrusion detection system, how to identify harmful information from huge amounts of data, to prevent and contain potentially dangerous behavior, the maintenance of the security and stability of network data transmission, to promote the healthy development of the Internet industry, has the significance of reality. In order to detect packet payload of the risk model, need to finish the matching of regular expression online. Although deterministic finite state machine(DFAs) to allow this operation in linear time, they might need higher memory requirements. in the memory, DFA storage overhead, mainly because of storing the state transition table storage, line width corresponding to the number of state DFA table, the table column width corresponds to the number of each state transfer edge |Σ|(Σ is the input characters alphabet).For grouping regular expression rules set is a kind of usage for the important method to solve the problem of DFA state expansion. So far, for the DFA in memory storage overhead the solution of a big problem, can be divided into two kinds, namely to reduce the number of state DFA and compressed transfer of DFA, through regular expression rules set grouping algorithm to compress the DFA storage space belongs to the solution of the above the first one. In this paper on the state machine optimization technology, on the basis of thorough analysis and research, the regular expression grouping algorithm is improved and it can be dealt with in the larger set of rules, improve the matching performance of state machine. The experimental data shows that the improved algorithm not only on the group performance is better than the average of grouping algorithm, it needs to update time is lesser also.In this article, some state machine optimization technique is adopted. One of DFA compression technology can make when dealing with N long string, up to 2 N state traversal. Moreover, the algorithm has lower complexity and very suitable for some of the dynamic scenes.As those with linear velocity to achieve the regular expression matching technology has been found. Different solution: such as automation types(type: determine vs non-deterministic) and implement the specific parameter configuration. Although each solution concentration in some specific set of rules and distribution model running is very good, but not for a large number of solutions, set of rules and circulation pattern sets system comparison. Therefore, for an implementer, in a large number of the existing algorithm and architecture scheme to make a right decision, it will be a great challenge.In order to solve this problem, this paper on a large number of regular pattern matching technologies to make a comprehensive assessment, consider two aspects of the algorithm and architecture. Also provides detailed performance report and through based on a simple evaluation of the rule set, determine the best configuration of clear guidance. These instructions can realize the actual regular matching system provide significant help.