Feature database generation method and device and corresponding traffic sorting method and device

Abstract

The invention discloses a method for generating a feature database. The method includes: grabbing target traffic corresponding to a target application; performing preliminary classification on the target traffic in a preset dimension; performing feature extraction on the classified target traffic, Match the extracted features with the saved historical feature database to predict the function of the target traffic; use each prediction result as a branch to generate an analysis tree for the target traffic; according to the Parse the tree to generate a new feature database. The present invention also provides a feature database generation device and a traffic sorting method and device using the feature database. The feature database generation method and device, traffic sorting method and device provided by the present invention can perform feature matching during traffic sorting according to the feature database, and optimize traffic matching accuracy and efficiency.

Description

technical field [0001] The present invention relates to the technical field of network optimization and traffic identification, in particular to a method and device for generating a feature database, and a method and device for sorting traffic. Background technique [0002] With the great popularization of the Internet in recent years and the diversification of the application market, the personalized needs of users have also increased. This requires the network optimization platform to have the ability to operate in a refined manner, and to manipulate, adjust, and optimize network resources in a smaller granularity. At present, a large number of services of different natures are encapsulated on the standard IPV4 protocol, and the preliminary traffic identification only relying on the third and fourth layer protocols is far from meeting the demand. Deep Packet Inspection (DPI) equipment is used to further identify Features over four floors. The common access mode of DPI equ...

AI Technical Summary

Problems solved by technology

[0003] However, this method has the following disadvantages: In the above identification process, for the DPI device operating on the network optimization platform, it is the passive receiver of traffic, that is, real-time analysis of traffic when it arrives, and the system's hardware performance and internal Logical efficiency puts high demands on

Moreover, the future trend of network traffic will inevitably develop in the direction of greater bandwidth demand. In the case of heavy traffic, feature matching will consume high system performance and may affect the quality of online services.

Images