A character string matching method, system, storage medium and device

Abstract

The present invention proposes a character string matching method, comprising: firstly, constructing a global migration table to store migration edges of the same input character with the same destination state; secondly, constructing a local migration table for each state, and storing Migration edges with different destination states in the global migration table, and use bitmap to further compress the local migration edge table. The construction time complexity of the global migration table is O(M×N), M represents the total number of DFA states, and N represents the number of unique characters in the alphabet, so the construction time of RDFA is less than that of existing algorithms; at the same time, the global migration table reduces With a large number of redundant migration edges, RDFA significantly compresses the DFA storage space; for each character read, RDFA only needs to look up the local migration table and the global migration table in the current state, thereby improving string matching throughput.

Description

technical field [0001] The invention relates to the fields of character string matching algorithms, deep data packet detection, application traffic identification and network security, and in particular to a ranking-based character string matching method, system, storage medium and device. Background technique [0002] String matching algorithms are widely used in network devices based on deep packet inspection, such as intrusion detection systems, traffic monitoring, etc. The algorithm finds out all matching signature string rules by matching the data packet content with a signature string rule set. With the increasing number of characteristic string rules, the string matching algorithm has become the performance bottleneck of network devices, and it is difficult to meet the performance and scalability requirements of deep packet inspection. [0003] The Aho-Corasick algorithm (AC algorithm) is currently the most widely used multi-string matching algorithm. It uses a deter...

AI Technical Summary

Problems solved by technology

When the number of string rules increases, the total number of DFA states (M) increases explosively, resulting in a rapid increase in DFA storage space overhead, exceeding the existing fast memory capacity of network devices

Although the existing string matching compression algorithms D2FA and ΔFA reduce the DFA space overhead, they have the following problems: the time complexity of D2FA to calculate the default migration edge is O(M2logM), and it is difficult to deal with DFA of large-scale string sets. In , reading a character in each state may need to find multiple default migration edges, resulting in low matching throughput; ΔFA needs to update the migration table of the next state once for each state transition, resulting in low string matching throughput

Therefore, existing string matching algorithms cannot meet the requirements of small string matching storage space, short construction time and high matching throughput at the same time

Images