面向多模态网络的可编程数据处理方法及系统设计

Programmable data processing method and system design for polymorphic network

针对多模态网络面临的兼容多种路由标识、网络协议和网络应用数据处理的问题,提出了一种兼容多模态的可编程数据处理方法。通过模态识别与分流模块识别不同网络模态的数据包,利用数据类型、偏移和长度表示多种类型的数据,并利用可选择匹配算法和处理动作的匹配动作数据表组成模态数据处理管道,处理不同类型网络的数据。利用所提方法,实现并验证了一个多模态网络数据处理原型系统。实验结果表明,模态数据处理管道可以隔离不同模态的网络并保证性能独立,匹配算法的类型、处理数据的长度会影响系统转发性能,而与数据类型无关。

Aiming at the problems of compatible with multiple route identification,network protocols and network application data processing faced by multimodal networks,a programmable data processing method compatible with multiple modes was proposed.Through the mode recognition and classify module to identify packets of different network modes,the data type,offset and length were used to represent multiple types of data,and the modal data processing pipeline was formed by using the matching action data table with selectable matching algorithm and processing action to process data of different types of network.A prototype system for data processing of polymorphic network was implemented and verified by using the proposed method.Experimental results show that the modal data processing pipeline can isolate the network of different modes and ensure the performance independence,the type of matching algorithm and the length of the processed data will affect the forwarding performance of the system,while the type of data will not.