新一代线速网络关键问题的研究

THE KEY RESEARCH OF NEXT-GENERATION WIRE SPEED PACKET CLASSIFICATION NETWORK

随着因特网的高速发展 ,因特网主干需要具备多业务能力的G位、甚至T位路由器。这就使得数据包输入处理成为主干路由器的瓶颈。本文分析了这一问题传统的纯软件算法解决方案 ,指出了其存在的历史局限性 :传统的算法只适用于具有大内存和比较强的处理能力的计算机 ;进而本文提出了应着眼于设计新的硬件体系结构来解决这一问题 ,最后本文总结提出了三种非软件的解决方法和思路 :发展新的CAM和TCAM设计工艺、基于人工神经网络的解决方法、基于协处理器的解决方法。

With the development of Internet,there are some new requirements to the kernel equipment of backbone router.In general,packet classification on multiple fields is a difficult problem.There are a lot of approaches to packet classification,which can be dealt in the software easily.But conventional soft algorithms have high time complexity and thus not used.Due to this the hardware based algorithms are now being implemented which is suitable for different applications.Our solution emphasis in developing new architectures such as CAMS and TCAMS,the neural networks based solution(LVQ),or the co processor based solution.