Online estimation method for traffic matrix of SDN based data center interconnection network

Abstract

The invention relates to the field of network traffic measurement and estimation of the SDN, and provides an online estimation method for a traffic matrix of an SDN based data center interconnection (SDN-DCI) network which can be used for simply, rapidly and accurately estimating the traffic matrix of the SDN-DCI large-scale network in an online mode. The method comprises the following steps: firstly, rapidly determining k OD flows with relatively large traffic of the current network by using a modified generalized gravity model; uniformly issuing corresponding flow tables according to an SDN flow table without changing the original route so as to measure the k OD flows with relatively large traffic, guarantee the estimation accuracy of the OD flows and ensure the load balancing of the device node flow table; and then positioning a link L with relatively large estimation errors according to the relative error between the link load measurement information of an SNMP protocol and the estimation value; and finally, defining the error factors of the OD flows, finding out n OD flows that have larger influence on the estimation error from a positioned link set L with relatively large estimation errors, and issuing the corresponding SDN flow table to measure the flow value so as to ensure that the estimation error of the traffic matrix can be reduced to a maximum extent.

Description

technical field [0001] The invention belongs to the field of SDN network traffic measurement and estimation, and in particular relates to an online traffic matrix estimation method of an SDN-based data center interconnection network (SDN-DCI). Background technique [0002] The continuous development of the Internet beyond expectations has resulted in rapid growth of network traffic. In order to improve the utilization of network resources, network managers need to establish traffic engineering to dynamically analyze and predict network traffic and plan efficiently. The traffic matrix represents the traffic flowing between all origin-destination (OD) node pairs in the network, and describes the specific distribution of traffic in the entire network. Traffic matrix, as the key input information of traffic engineering, is an important parameter of network performance and provides an important reference for rational planning of network routing. Due to the importance of the flo...

AI Technical Summary

Problems solved by technology

However, due to the limitation of hardware resources, the entries of flow entries are limited, and it is impossible to measure all OD flows

For the measurement of online traffic matrix, how to distribute the flow table to each network device in a balanced manner under the condition of limited hardware resources and satisfying routing constraints, and how to quickly and accurately give the estimated value of the traffic matrix is ​​still not very good. s solution

Images