Application-based high-performance refined shunting method

Abstract

The invention discloses an application-based refined shunting system, and the system comprises a basic shunting module which is used for receiving and identifying messages, including control messages and data messages, maintaining a user context information base according to the control messages, carrying out matching association on the data message and the user context information base, marking the data message, carrying out basic classification rule search on the marked data message, and carrying out shunting judgment on the message returned by the fine classification module and executing a corresponding action; and a refined classification module which is used for classifying the first N packets of the corresponding flow of the data message which does not hit the basic classification rule, wherein the original data message and the classification result are combined into a message returned by the refined classification module. The data message is marked through the user information, can be accurately shunted according to the user information, can be added into refined shunting, can be shunted based on application, can unload low-value traffic through refined shunting, reduces the performance pressure of a back-end service system, and achieves the target of high-performance refined shunting.

Description

technical field [0001] The invention relates to network communication distribution, in particular to an application-based high-performance fine distribution method. Background technique [0002] Mobile Internet traffic is increasing day by day, and the annual traffic of mobile Internet system access has increased by 50% to 100%. The existing analysis system can no longer handle the full traffic in consideration of factors such as space, energy consumption, and funds. The current front-end Devices can only be classified according to the IP layer. The ability to identify and screen is weak, the distribution granularity is not precise, and the ability to screen and identify data by the application layer is lacking, which will cause a large amount of low-value traffic to flow to the back-end system. [0003] In the prior art, the patent CN109768936A discloses a refined distribution system and distribution method, which makes up for the coarse-grained defects of classification ba...

AI Technical Summary

Problems solved by technology

[0002] Mobile Internet traffic is increasing day by day, and the annual traffic of mobile Internet system access has increased by 50% to 100%. The existing analysis system can no longer handle the full traffic in consideration of factors such as space, energy consumption, and funds. The current front-end Devices can only be classified according to the IP layer. The identification and screening capabilities are weak, the distribution granularity is not precise, and the lack of the ability to filter and identify data by the application layer will cause a large amount of low-value traffic to flow to the back-end system.

[0003] In the prior art, the patent CN109768936A discloses a refined distribution system and distribution method, which makes up for the coarse-grained defects of classification based on the IP layer, but there are still two problems in this patent: first, the method is fine The degree of modernization is not enough, and it does not have the ability to classify and trace the information of mobile Internet end users

Images