Abstract: Modified Aho-Corasick（MAC） algorithm，an effective string matching algorithm with advantages of both compact memory and high performance，is proposed.By employing the characteristics of same states observed from the deterministic finite state automata，the proposed MAC significantly reduces the memory requirement without sacrificing high speed.The MAC algorithm also provides high flexibility that it can be tuned to fit specific performance requirement and resource constraints.The experimental results show that the performance of ACMS is over 1.51~2.40 times in software implementation compared to existing state-of-the-art algorithms.

Key words: Modified Aho-Corasick（MAC） algorithm, Network Intrusion Detection System（NIDS）, pattern matching, Deterministic Finite-state Automata（DFA）, Nondeterministic Finite-state Automata（NFA）

摘要： 提出一种高性能的模式匹配算法——MAC算法，它通过使用从确定性有限状态机（DFA）中得到的特征等同态，在保证高速匹配的前提下，极大地减少了内存需求。同时，该算法具有高度的灵活性，即通过调整就可以适应不同的特定性能和资源限制的要求。在软件使用环境中的实验结果表明，MAC算法的内存使用性能相对目前先进的模式匹配算法提高了1.51~2.40倍。

关键词: MAC算法, 网络入侵检测系统, 模式匹配, 确定性有限状态机, 非确定性有限状态机