Service perceived flow policy customization method based on software defined network

Abstract

The invention belongs to the technical field of traffic scheduling and discloses a service perceived flow policy customization method based on a software defined network. Traffic recognition and traffic scheduling are carried out in a controller. Through expansion of basic modules of the controller, two modules: a service traffic recognition and classification module and a service traffic policy scheduling module are developed and are respectively used for processing traffic classification and scheduling functions. The method is a relatively reasonable traffic recognition and traffic scheduling scheme in the software defined network (SDN). Service perception can be carried out on the traffic. The service traffic can be recognized and classified relatively precisely. Forwarding policies ofdifferent QoS levels are made for the service traffic of different classes. The quality of service required by business access provided by an enterprise portal for a user is guaranteed. The QoE of theuser for the service is improved. Moreover, the load balance and a utilization rate of a link are guaranteed to a certain degree.

Description

technical field [0001] The invention belongs to the technical field of traffic scheduling, and in particular relates to a flow policy customization method based on software-defined network service perception. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: The traffic scheduling problem is a kind of traffic engineering problem in the traditional network. In the SDN software-defined network, the traffic scheduling is usually called flow scheduling. The traffic scheduling strategy is not completely unified, and the same flow scheduling strategy is used for most traffic, such as load balancing and shortest path. But often, there are many types of traffic that need to be treated differently to meet the high-quality transmission requirements required by users (such as key application and multimedia application traffic, etc.). For this part of traffic, it is called service flow, which needs to be customized Specia...

AI Technical Summary

Problems solved by technology

[0003] In summary, the problems in the prior art are: The existing traffic scheduling strategy cannot identify the flows generated by new applications in the network, so it is difficult to accurately provide high-quality services for all flows, and satisfy multiple QoS (quality of service) constraints on the classified service flows When scheduling at the same level, the topology and link state information of the entire network are not fully considered, and the load balance of the entire network link cannot be guaranteed. Some links may have a high load while a certain link has a low load and are often idle, etc. Unbalanced load problem

Images