SLA-Driven Radio Resource Management Using Control Parameter Optimization in 5G Network Slicing

Abstract

In traditional telecommunications networks, introducing new functions or processes, especially with diverse approaches to quality assurance, often requires fundamental changes to the architecture and configuration of existing networks, which will lead to hardware modifications. Consequently, such changes usually faced widespread resistance from operators due to the significant costs involved. A viable solution that gained widespread popularity in the late first decade of this century is the software-centric redesign of network functions to facilitate centralized management, reduce expansion and update costs, and enhance the overall efficiency of the system. Alongside extensive academic research in this direction, this solution is now significantly implemented in the industry, particularly in fifth-generation communications and beyond. The diverse and potentially conflicting requirements of new applications in modern wireless telecommunications present a significant challenge in maintaining connectivity and coherence among various blocks of an integrated network to meet the quality needs of diverse applications. One proposed solution in this field, based on the approach of software-defined networks, is network slicing. The main concept of slicing involves sharing various physical network resources over virtual networks with centralized management and predefined quality of service requirements. For this purpose, different network slices must be managed and configured by a central control unit. In the cellular wireless network studied in this research, this unit is introduced as the mapping layer. This layer monitors its serviced network and manages the allocation of radio resources to slices based on a specific method to meet the service needs of each slice. Following an initial introduction, the proposed idea is compared with some existing and new methods. Simulation results from this research indicate that the proposed slicing method performs better in terms of meeting key performance indicators compared to other methods reviewed, especially when the demand for resources exceeds the capacity allocated to a slice, relative to the agreed service level.