Core Network Slicing Resource Management Model Based on Markov Decision Process and Fairness Resources.

The fifth-generation cellular network can handle extremely diverse services and user requirements including high transmission rates, low latency, and large connections. The concept of network slicing appears as an effective solution to satisfy these various needs, where different virtual networks for various service scenarios are implemented on the same physical infrastructure. The sharing technique has several issues, such as ensuring Quality of Service (QoS) satisfaction, fairness, and performance resource allocation across different slices. The variability of these slices mostly resides in their number of demands and the priority of each slice. Slices that have high demands and priority are often allocated a significant number of resources. However, due to the limitation of resources management strategies of network slicing. Moreover, the resource allocation strategy excludes the slice requests of low priority with its demands from the slice. In this scenario, some slices with low demand and priority may be suffering from starvation. The primary focus of this study is to address and overcome these issues in core network slicing. This paper proposes a resource allocation management model Based on Markov decision process and adaptive fairness resources, that considers the Quality of service of each slice based on the changing demands per slice at a specific decision epoch. The experiment results show that the proposed solution with maximizing utilization policy has outperformed by 40% compared to exist scheme in terms of QoS, which proposed a prioritized dynamic allocation scheme. Moreover, fair resource allocation ensures maximizing the overall network while maintaining the requirements of each slice and QoS guarantees. [ABSTRACT FROM AUTHOR]