AI Based Resource Management for 5G Network Slicing: History, Use Cases, and Research Directions

• Published in: Concurrency and computation Vol. 37; no. 2
• Main Authors: Dubey, Monika, Singh, Ashutosh Kumar, Mishra, Richa
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 25.01.2025; Wiley Subscription Services, Inc
• Subjects: 5G mobile communication; Artificial intelligence; Communications traffic; Edge computing; Industrial applications; Network function virtualization; Network slicing; Quality of service architectures; Resource management; Resource scheduling; SDN; Software-defined networking; State-of-the-art reviews; Wireless networks
• ISSN: 1532-0626; 1532-0634
• DOI: 10.1002/cpe.8327

ABSTRACT 5G, 6G, and beyond networks promise to support vertical industrial services with strict QoS parameters, but the hardware‐based "one‐size‐fits‐all" model of legacy networks lacks the flexibility needed for diverse services. The foundation of 5G networks lies in softwarization, with network slicing, Software Defined Networking (SDN), and Network Function Virtualisation (NFV) serving as its core components. The network‐slicing‐based shared network environment necessitates an intelligent and flexible resource management approach. In this case, traditional approaches are no longer suitable for dealing with a dynamic network environment. With recent advancements, AI‐based approaches have the potential to manage resources autonomously. This paradigm shift underscores the need for deep and extensive investigation. However, existing literature on this subject is fragmented and lacks a cohesive overview of network slicing. To address these gaps, our review paper aims to provide a comprehensive scope of network slicing in a unified manner. In this sequence at first, this paper presented a conceptual overview of network slicing and enabling technologies, including SDN, NFV, and edge computing. Secondly, this paper identifies the relevant phases of resource management and presents AI‐based resource management for network traffic classification, admission, allocation, and scheduling. Finally, it also discusses the deployment of network slicing‐enabled key use cases and their practical deployment, the research gap, and open research challenges. To the best of our knowledge, this is the first attempt to critically analyze and present a consolidated review of the state of the art in network slicing resource management modules and network slicing‐enabled key industrial use cases. This paper aims to guide researchers in developing innovative solutions and assist network players in the practical deployment of network slices for industrial applications.