Parallel network packet classification method and system based on multi-core processor and network equipment

Abstract

The invention provides a parallel network packet classification method and system based on a multi-core processor and network equipment, and the method comprises the steps: receiving a network data packet, putting the network data packet into a queue, and preprocessing the network data packet; and simultaneously classifying the queued network data packets one by one by adopting the same matching process in an active circulation mode by a plurality of cores of a multi-core processor, and outputting a matching classification result of the network data packets according to an input sequence of the network data packets. The parallel network packet classification method based on the multi-core processor is proposed so that an input sequence and an output sequence of network data packets are consistent, avoid disorder, eliminate the competitive overhead among multiple cores and give full play to the performance advantages of multi-core parallel. Non-competitive intra-packet level parallel multi-core network packet classification can be realized, the multi-core parallel efficiency advantage of the multi-core processor is fully exerted, and the network packet classification process is accelerated.

Description

technical field [0001] The invention relates to the field of network communication, in particular to a multi-core processor-based parallel network packet classification method, system and network equipment. Background technique [0002] The network packet classification technology refers to matching the network packets in the received data stream with the rule set in the classifier, and performing corresponding operations on the network packets according to the action type corresponding to the matched rules. A classifier is a collection of rules or policies. Each classification rule specifies a category or a specific flow. By comparing and matching the specific fields in the header of the received network packet with the corresponding fields in the rule set, the classifier can be judged. Which rules the packet meets, and then determine the category or network flow it belongs to. Packet classification functions are required for many network services, such as routing, firewal...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the problem that when the current software network packet classification method is parallelized on a multi-core processor, the output of the packet-level parallel mode is inconsistent with the input order and the problem that the parallel effect is not ideal due to a large amount of multi-core competition overhead of the parallel mode within the package level , a parallel network packet classification method based on multi-core processors is proposed. This method aims at searching and matching network packets based on rule sets, and decomposes multiple domains of the rule set into multiple dimensions. When performing rule matching and classification on network data packets , first perform a search on each dimension separately to obtain the sub-results of each dimension, and then combine the search results of all dimensions to obtain the final matching result

Images