一种基于AC算法的电力业务快速识别方法

电力公司主要业务系统已逐步采用移动作业终端通过公用信息网络接入电力信息内网,业务系统在内外网数据交换过程中通过协议特征进行业务识别。针对业务识别过程存在匹配时间长、内存消耗大等问题,提出了一种基于AC（Aho-Corasick）算法的电力业务快速识别方法。在分析电力信息网络安全接入流程、高密度流量的协议特征、安全网关性能瓶颈的基础上,提出基于传输层协议分组的并行处理方式,提高匹配处理的吞吐量;采用基于位图的存储方式,降低构建和存储协议特征规则库所需的内存空间;最终通过算例证明了该方法能提升协议匹配速率,降低协议特征规则库的内存消耗,使协议特征识别过程的时间开销和空间开销达到最优平衡。仿真结果验证了所提方法的有效性和准确性。

A Quick Identify Method for Electricity Business Based on AC Algorithm

Most business systems in electric power company use mobile terminals to access internal electricity information network, and signatures of protocols are used to identify different businesses while exchanging data between internal and external network. In order to solve problems of slow pattern matching and large memory consumption, a quick electricity business identify method based on AC algorithm is proposed. After analyzing security access procedure in electricity information network, protocol features of high-density traffic and security gateway performance bottlenecks, the proposed method groups signatures of protocols according to the transport layer protocols for parallel processing to increase matching throughput. The bitmap-based storage is selected to reduce memory consumption of rules library. At last, simulation results show that proposed method can increase matching throughput and reduce memory consumption.