Computing and Storage Resources Allocation of UPF Based on Isolation in Private 5G Networks

The private 5th generation mobile networks (5G) have emerged to provide customized services to the vertical industries, where the resources of the radio access network (RAN) and the core network (CN) can be shared among the private 5G networks in part or fully [1]. Therefore, it is significant to achieve resource isolation to guarantee the quality of service (QoS) for each private 5G network while other networks vary. Since user plane function (UPF) is an important network function of 5G CN, the resource isolation should be achieved when assigning the resources to the UPFs. However, the resource isolation performance of UPFs is ignored in the present researches. Therefore, this paper studies the computing and storage resources allocation of UPFs based on the isolation performance in the private 5G networks. First, the isolation index of one UPF is defined as the weighted sum of the performances for the multiple QoS requirements to represent the resource isolation performance. Second, the resource demands of one UPF are obtained to minimize the cost and to meet the isolation requirement. To achieve this, an optimal problem is formulated and a particle swarm algorithm is designed. Finally, the severs and the corresponding computing and storage resources are assigned to multiple UPFs, where the resource utilization of the system is maximized and the resource demands of the UPFs are met. To this end, the other optimal problem is formulated and solved by providing a greedy algorithm. By simulating, it can be seen that the designed particle swarm algorithm can obtain the resource demands with minimum cost for one UPF when the isolation requirement is met. And the proposed greedy algorithm can allocate enough resources to the UPFs and achieve high resource utilization of the system without increasing running time.