User association and resource allocation in 5G (AURA-5G): A joint optimization framework

摘要

5G wireless networks will be extremely dense, given the projected increase in the number of users and access points (APs), as well as heterogeneous, given the different types of APs and the applications being accessed by the users. In such challenging environments, efficient mobility management (MM), and specifically user association, will be critical to assist the 5G networks in provisioning the Quality of Service (QoS) of diverse applications 5G targets to serve. Whilst determining the most suitable AP for the users, multiple constraints such as available backhaul capacity, link latency, etc., will need to be accommodated for. Hence, to provide an optimal user association solution, in this paper we present a joint optimization framework, namely AURA-5G. Under this framework we formulate our user association strategy as a Mixed Integer Linear Program (MILP) that aims to maximize the total sum rate of the network whilst optimizing the bandwidth assignment and access point selection. We analyze multiple active application profiles simultaneously, i.e. enhanced Mobile Broadband (eMBB) and massive Machine Type Communication (mMTC), in the network and study the performance of AURA-5G. Additionally, we provision a novel study on the multiple dual connectivity modes, wherein the user can be connected to either one macro cell and a possible small cell, or with any two favorable candidate access points. Utilizing the AURA-5G framework, we perform a novel comparative study of all the considered scenarios on the basis of total network throughput, performance against baseline scenario and system fairness. We show that the AURA-5G optimal solutions improve the performance of different network scenarios in terms of total network throughput (by 38x-690x) and system fairness as compared to the baseline scenario, which is a conventional user association solution. Further, we also present a fidelity analysis of the AURA-5G framework based on the backhaul utilization, latency compliance, convergence time distribution and solvability. And since, a given network cannot always guarantee to satisfy the future network loads and application constraints, we show that AURA-5G can be utilized by the operators/vendors to evaluate the myriad network re-dimensioning approaches for attaining a feasible and optimal solution. Henceforth, we then explore the possibility of network re-dimensioning and study its impact on system performance for scenarios where the performance of AURA-5G is severely impacted due to the extremely strict nature of the constraints imposed in the MILP.