Machine Learning-based Slice allocation Algorithms in 5G Networks

One of the primary aims for 5G networks is the flexibility to support vertical services as the primary requirement. The use cases of 5G networks are AR/VR, eHealth, Video Streaming, automation works, automated vehicles, etc. All use cases are associated with its requirements like availability, reliability, latency, and throughput in terms of Quality of Services (QoS) and Quality of Experience (QoE). In 5G, the Network-Functions Virtualization (NFV) and Software-Defined Networking (SDN) combined to provide the concept of Network Slicing. Network Slicing runs multiple logical networks as virtually independent operations over shared physical infrastructure. In this work, we address the slice allocation problem using machine learning algorithms with the dataset. The performance evaluation of the slice allocation algorithm has addressed using different machine learning models. In the simulation, we demonstrated that our proposed method allocates the best slice for service with high accuracy.

I. Introduction

Network Slicing concept was first conceived by a standardization body, NGMN [5] later that considered being the key enablers in the next fifth-generation (5G) networking. NFV has emerged with possible approaches to support for an open network [1], it allows service providers to become more flexible. The NFV architecture supports bottom-up hierarchy to support vertical services and the combination of SDN, NFV enables trends for the network softwarization [2]. 5G networks aim to achieve various QoS in a possible network scenario in terms of latency and throughput [3].