A vnf migration method based on federated learning two-way GRU resource demand prediction

Abstract

The invention relates to a VNF migration method based on federated learning bidirectional GRU resource demand prediction, belonging to the technical field of mobile communication. The method includes: S1: In the network slicing scenario, considering the VNF migration problem caused by time-varying network traffic and the VNF migration delay problem caused by the lack of prediction of VNF resource requirements, the FedBi‑GRU algorithm is used to predict the resource requirements of the VNF; S2 : Calculate the resource utilization rate of physical nodes according to the resource demand prediction results, and determine the physical nodes with overloaded resources or lightly loaded resources in the network system. Through VNF migration, system energy consumption optimization and load balancing can be realized while ensuring network performance; S3: Using the deep reinforcement learning method of DPPO to obtain the optimal decision of VNF migration. The invention can reduce the migration times of the virtual network function and reduce the energy consumption of the network system, and can ensure the load balance of the network system.

Description

technical field [0001] The invention belongs to the technical field of mobile communication, and relates to a VNF migration method based on federated learning bidirectional GRU resource demand prediction. Background technique [0002] Software Defined Networking (SDN) and Network Functions Virtualization (NFV) are two new architectures for managing network systems, the main idea behind SDN is to make the network directly programmable and separate the control plane from the data plane to provide a centralized view for the network , to provide manageability for complex network systems, NFV converts physical layer resources into virtual resources, separates software instances from the underlying dedicated hardware, and makes the network flexible. [0003] Using NFV technology, traditional network hardware resources can be virtualized into multiple virtual machines and various network element function software of the operator network can be instantiated into virtual network func...

AI Technical Summary

Problems solved by technology

[0004] Existing VNF migration problems are based on real-time network information for VNF migration. There are few literatures on VNF resource demand forecasting to formulate VNF migration plans in advance, and the prediction methods used in these literatures are all centralized machine learning.

In addition, in the face of the high-dimensional and complex VNF migration remapping space, the existing heuristic algorithm is not enough to find the optimal VNF migration scheme

Images