Interference and QoS-Aware Resource Allocation Considering DAS Behavior for C-RAN Power Minimization

Abstract

This work aims to minimize the power consumption of cloud virtualization components by addressing the important challenge of designing allocation schemes that cater for both radio and computational resources in a virtualized cloud-based radio environment. Unlike previous efforts, we consider the realistic behavior of radio resource heads (RRHs) associated with one baseband unit (BBU), acting as a DAS. We first formulate the admission control (AC) and the RRH–BBU mapping problem, subject to constraints on user throughput requirements, computational capacity in the BBU pool, and according to the assumption of DAS behavior. As the optimal solution is practically intractable for large-scale dynamic networks, we propose a two-level resource allocation framework, based on two developed algorithms: a first one for AC at the radio level and a second one for RRH–BBU mapping at the computational level. The two algorithms are designed to allow both the levels to consider each other’s constraints and particularities. Results obtained from an extensive simulation-based performance evaluation show the high performance of our proposal in terms of radio-related metrics, number of accepted users, power saving, and energy and spectrum efficiency, compared with three baseline reference schemes.