A method and system for packet classification based on convolutional neural network

Abstract

The invention provides a data packet classification method and system based on a convolutional neural network. The method includes merging each rule set in the training rule set to form multiple merging schemes, determining the optimal merging scheme for each rule set in the training rule set based on performance evaluation; combining each rule set in the training rule set and the target rule set The distribution of the prefix combination is converted into an image, and the convolutional neural network model is trained with the image and the corresponding optimal merging scheme; the target rule set is classified based on the image similarity, and the corresponding hash table is constructed for packet classification. The method of the invention can remarkably improve the data packet search performance, increase the data packet search speed, and increase the update speed of the rules. Through the mutual cooperation between the online system and the offline system, the system of the present invention can ensure that the online system realizes efficient search of data packets and rapid update of the rule set, and can monitor the update of the rule set to always reflect the latest state of the network.

Description

technical field [0001] The invention relates to the technical field of searching and classifying data packets in a computer network, in particular to a method for classifying data packets based on a convolutional neural network. Background technique [0002] Packet search and classification implements packet classification processing through predefined or dynamically generated rule sets, and is a basic key function in switches, routers, firewalls, load balancers, cloud platform software switches, and other network devices. Scenarios such as software-defined networking, network function virtualization, and cloud computing require frequent rule updates. Rules are stored in a data structure such as a decision tree to achieve high-speed matching and forwarding of data packets, but the update speed of the rules is slow. When the rules are updated, the speed of data packet matching and forwarding will be greatly reduced. The hash-based packet classification method can support fas...

AI Technical Summary

Problems solved by technology

Although existing methods such as Pruned Tuple Space Search, TupleMerge, and PartitionSort improve the TSS search speed by sacrificing the update performance, they still cannot support efficient classification of data packets and high-speed online rule update at the same time.

Images