Threshold-based scaling model for NSSI blocks in 5G network slicing

Abstract

In the 5G core network, network functions may be dynamically scaled out and in to modify capacity for network slices. Automatic scaling enhances performance by scaling-out network function instances while decreasing operational expenses by scaling-in instances. 5G network functionalities must be deployed and stopped concurrently across numerous NFI instances, which are commonly treated as a block of NSSI instances. As a result, determining the number of instances that must be deployed concurrently, as well as the number of instances that must be shut off when no longer in use, will have a substantial impact on the system's cost efficiency and QoS service quality. Furthermore, using frequent block setup may drastically lower system performance (setup cost), hence the problem of reserving (presetting) specific blocks must be considered. In this article, we will use the Markov queuing model to represent the whole system for scaling NSSI instance blocks corresponding to network slices in the 5G core network. The model will include two thresholds for the setup/removal (scaling-out/scaling-in) of NSSI blocks based on system user traffic. The numerical analysis findings suggest that the scaling model of NSSI blocks based on the threshold presented in this study can efficiently control system resource allocation.