论文部分内容阅读
在基于有限状态自动机的正则表达式匹配技术中,存储需求和匹配性能是一对相互制约的因素.统计分析发现,对于给定的自动机,状态的访问概率分布呈Zipf规律.为解决存储和性能的矛盾,设计并实现了基于Net Magic的两级存储的匹配引擎;根据状态的访问概率特性,将访问概率较高的状态配置在小容量的高速存储器中,访问概率较低的状态配置在大容量的低速存储器中,从而达到存储与性能的平衡.同时利用FPGA内部多RAM块特性,实例化多个匹配引擎,进一步使匹配速度线性提升.在资源充裕的条件下,理论上可达到65Gbps的吞吐量.实验表明单位存储代价大幅降低,但受限于Net Magic存储资源、频率及端口速率,实际性能为2.1Gbps.
In regular expression matching based on finite state automaton, storage requirement and matching performance are a pair of mutual restraint factors.Statistical analysis shows that, for a given automaton, the access probability distribution of the state is Zipf rule.In order to solve the problem of storage And the performance of the conflict, the design and implementation of a two-level storage based on Net Magic matching engine; according to the state of access probability characteristics, the access probability of the state is configured in a small capacity of high-speed memory access to low probability of configuration In large-capacity low-speed memory, so as to achieve the balance of storage and performance.At the same time the use of multi-RAM block features within the FPGA instantiation of multiple matching engine to further improve the linear matching speed.Under the conditions of sufficient resources, theoretically achievable 65Gbps throughput.Experiments show that the unit storage cost significantly reduced, but limited by Net Magic storage resources, frequency and port speed, the actual performance of 2.1Gbps.