Deep learning-based SDN network traffic advantageous monitoring node dynamic selection system and dynamic selection method thereof

Abstract

The invention discloses a deep learning-based SDN network traffic advantageous monitoring node dynamic selection system and a dynamic selection method thereof. A SDN control plane comprises a forwarding calculation module, a node updating module and a path prediction module. A SDN data layer comprises a network resource module. The dynamic selection method includes an advantageous monitoring node pre-screening stage and an advantageous monitoring node dynamic updating stage. The advantageous monitoring node pre-screening stage only operates when the system is cold boosted. The advantageous monitoring node dynamic updating stage operates adaptively in a closed-loop self-feedback mode after the system startup is completed. The dynamic selection system, in view of selecting a monitoring node, preferentially selects a switch with the densest traffic traversal as a traffic monitoring node, and achieves a purpose of increasing traffic collection non-redundancy rate and reducing the traffic monitoring overhead while maximizing the traffic statistical information.

Description

technical field [0001] The invention belongs to the technical field of network security, and in particular relates to a dynamic selection system and a dynamic selection method for an SDN network flow advantage monitoring node based on deep learning. Background technique [0002] As a new network architecture, software-defined networking (SDN) realizes centralized management of network resources by decoupling the control plane and data plane of network devices. This method of separating the control rights of IP network devices and having the controller manage the network in a unified manner can shield the differences of the underlying heterogeneous networks and effectively reduce the dependence of network management on the underlying network devices. Although SDN has a significant effect on improving resource utilization efficiency and enhancing the refinement of network management, in order to realize this centralized management method, the SDN controller needs to obtain a g...

AI Technical Summary

Problems solved by technology

(1) In SDN traffic monitoring based on traffic sampling, the most discussed is to use sketches to allocate traffic monitoring resources and manage traffic monitoring tasks; Obtain all information about traffic changes, resulting in an incomplete and inaccurate global view of the network

(2) In Push-based SDN traffic monitoring, the SDN controller can passively receive the traffic statistics information pushed from the switch to grasp the active flow situation, and can obtain as much traffic change information as possible with less overhead under normal circumstances; but Because this type of method has strict requirements on the software and hardware of the switch, and will still generate large communication and computing overhead when the traffic changes frequently, it is rarely used at present.

[0005] Although Pull-type technical solutions are currently receiving the most attention, their implementation does not fully consider the fact that network traffic distribution generally follows the "20 / 80" rule, further reducing the redundancy rate of SDN traffic monitoring information and monitoring overhead faces bottlenecks

Images