Method and device for shunt of network flow

Abstract

The invention relates to a method and a device for shunt of network flow. The method comprises the following steps that attributive strategy grids of the current network flow are inquired in a strategy grid array according to one or more semantic parameters carried by the current network flow, wherein the strategy grid array is created with dimensionality and granularity of the network flow as an index, and the dimensionality and the granularity are divided based on the user properties under the current business scene; and shunt is executed for the current network flow according to shunt strategies corresponding to the attributive strategy grids, wherein the shunt strategy corresponding to each strategy grid in the strategy grid array is arranged to a global shunt strategy or one or more grid shunt strategies. According to the technical scheme of the invention, the network flow can be classified and labelled, so that each network flow source carrying the semantic parameter can be mapped into a unique grid of the strategy grid array and is subjected to shunt processing according to the shunt strategy corresponding to the attributive grid.

Description

technical field [0001] The present application relates to the technical field of the Internet, in particular to a method and device for distributing network traffic. Background technique [0002] With the increase of business volume, the rapid growth of access volume and data flow, each core part of the network has a corresponding increase in processing power and calculation intensity, making it impossible for a single device to bear it. In view of this situation, if the hardware upgrade is adopted to solve the problem, it will not only require higher cost investment than before, but also cause a waste of existing resources, and, no matter how excellent the performance of the equipment is, it cannot meet the demands of the ever-increasing business volume. Demand, so the load balancing mechanism came into being. [0003] Load balancing is to balance and distribute the load to multiple operating units, such as Web servers, FTP servers, enterprise key application servers, and ...

AI Technical Summary

Problems solved by technology

However, it is impossible to implement different load balancing strategies for different traffic groups, especially for the traffic of different user groups.

[0007] At the same time, most of the existing load balancing solutions are to formulate distribution strategies for the traffic itself. Based on the non-semantic distribution technology at each layer of the TCP / IP network protocol, it is aimed at networks that can have semantic parameters of business nature, such as user access traffic. For traffic, it is impossible to provide a semantic distribution ratio scheme. For example, it is impossible to provide a semantic and scenario-based distribution strategy for distribution ratio based on semantic parameters such as the user's affiliation, user wishes, and user preferences.

This shortcoming also makes the existing load balancing always adopt the implementation scheme directly deployed at the network layer, which cannot be closely related to the business decision system and implement semantic distribution decision-making

Images