A Slice-Centric and SLA-Aware Flexible Radio Resources Allocation Solution for 5G Network Slice Management

Radio Access Network (RAN) slicing is a rudimentary facet of 5G New Radio (5G NR) technology that promotes the concurrent support of diverse applications with distinct service level agreements (SLAs) on a common physical network infrastructure. SLAs delineate specific Quality of Service (QoS) requirements and key performance indicators (KPIs), including throughput, latency, and packet loss rate, to ensure reliable and efficient network performance (as defined by 3GPP in TS 28.530). However, effective radio resource allocation (RRA) in RAN slicing remains challenging due to dynamic wireless channel conditions, high user mobility patterns, and limited frequency spectrum. This paper introduces FNSM, an innovative slice-centric, SLA-aware, flexible radio resource allocation solution for advanced network slice management in 5G open-RAN (O-RAN) architecture. FNSM uses a smart two-stage resource allocation approach that considers immediate user requirements and service quality assurance to efficiently distribute resource block groups (RBGs) in resource-limited 3GPP-based RAN slicing scenarios. In comparison to existing RRA solutions, FNSM introduces a novel SLA-spectral efficiency measure (SEM) to allocate resources based on service compliance and spectral efficiency performance, facilitating SLA assurance while optimizing resource use. The low-complex FNSM solution systematically adapts to dynamic environments and enables efficient and scalable resource allocation without requiring extensive training or hyperparameter tuning. The comprehensive experimental evaluations using realistic spectral efficiency and channel conditions in a dense urban simulation environment with eMBB, URLLC, and BE slices demonstrate the superior performance of FNSM compared to state-of-the-art approaches.