Method and device for scheduling maximum probability path flow of SDN-oriented data center network

Abstract

The invention discloses a method and a device for scheduling a maximum probability path flow of an SDN-oriented data center network. The method comprises the following steps: S1: obtaining flow relevant information via a switch; S2: judging the flow and judging whether the flow is an elephant flow; if the flow is not the elephant flow, obtaining optional paths of the flow according to the information of the flow and global topological information of controller maintenance, and then selecting any path for the flow; and if the flow F is the elephant flow, selecting a probability path algorithm to calculate one path which meets flow bandwidth requirements for the flow F; and S3: after selecting one path for the flow F, sending a flow entry to the switch; and after the flow F is matched with the flow entry, carrying out data forwarding according to an action or matching residual data of the flow F to a new flow entry for forwarding; and the device is used for realizing above method. The method and the device disclosed by the invention have the advantages of simple principle, wide application range, easiness in realization, and capability of reducing network congestion to take a network balancing effect, etc.

Description

technical field [0001] The present invention mainly relates to the technical field of data center network, in particular to a maximum probability path flow scheduling method and device for SDN data center network. Background technique [0002] The ever-increasing data center network scale and business requirements pose severe challenges to data center network load balancing, and the emergence of a new network architecture based on SDN brings new ideas and opportunities to load balancing research. The controller in the SDN network can obtain the topology, nodes, links, and flow information of the global network, and then can perform centralized control and scheduling on the network flow. The decoupling and separation of the SDN network control plane and forwarding plane realizes the abstraction and high-speed forwarding of the underlying switch resources, and the northbound interface provided by the controller can facilitate developers to develop new functions and research in...

AI Technical Summary

Problems solved by technology

This algorithm does not fully consider the bandwidth usage of all paths that meet the flow bandwidth requirements, and may lead to fragmentation of the remaining bandwidth of the link, so it cannot achieve global dynamic load balancing well

[0011] The problems existing in the above-mentioned existing flow scheduling algorithms are: 1) the demand for flow bandwidth and the remaining bandwidth of the link are not fully considered; 2) the problem of flow scheduling is not considered in the long-term overall, and the flow scheduling cannot be reasonably used for multi-path; 3) flow scheduling exists Link Bandwidth Fragmentation Problem

Images