Design And Implementation Of Traffic Scheduling System For 5G Core Network

Recently,with the large-scale deployment and commercialization of the 5 G network,cellular data is growing rapidly and various novel network applications are emerging.The huge traffic load and the diverse service demands brought by novel network applications have put huge pressure on network operation and maintenance for mobile operators.How to reasonably schedule the traffic on the 5G Core Network User Plane to ensure the user experience has become a key issue for mobile operators.Therefore,this paper designs and implements a traffic scheduling system for 5G Core Network User Plane.This system uses multi-type,realtime,and high-precision network state information of 5G Core Network User Plane to reasonably schedule PDU session traffic and improve quality of service.Thus,the system is mainly divided into network state sensing subsystem and routing decision subsystem,and routing decision subsystem.The network state sensing subsystem senses various network state information of the 5G Core Network in real time through in-band network telemetry.The subsystem combines the forwarding of GTP-U packets and network measurement at the 5G Core Network User Plane and inserts network telemetry information into the Extension Header of GTP-U packets,so as to realize the network measurement along the road and ensure good compatibility.The routing decision subsystem is responsible for planning a reasonable routing path for the PDU sessions initiated by UEs.An improved ant colony optimization algorithm is designed and implemented in this paper for traffic scheduling(IACO-TS).In response to the problem that the original ant colony algorithm does not have enough initial search capability,the optimized ant colony algorithm in this paper improves the initialization of pheromone and proposes a pheromone update method based on the maximum-minimum ant colony system and elite ant strategy.In addition,the algorithm also introduces the network state information such as link bandwidth and delay of 5G Core Network User Plane into the heuristic function and pheromone calculation function,so as to realize the intelligent routing decision based on the current network state.In this paper,the traffic scheduling system is deployed on the open source 5G Core Network named Free5GC.Functional tests and performance analysis have been conducted on corresponding network topology.The experimental results show that the traffic scheduling system implemented in this paper can sense the link delay,real-time bandwidth of UPF network ports,UPF processing delay and other related network state information on the 5G Core Network User Plane in real time at a small cost.The system can reasonably schedule the UE-initiated PDU session traffic by the algorithm proposed in this paper and outperforms the traditional traffic scheduling methods in terms of performance parameters such as delay,throughput,and packet loss rate,improving the network service quality at the user-plane of the 5G core network.