Mapeamento otimizado entre RRH-BBU visando o balanceamento de carga em arquitetura C-RAN utilizando inteligência computacional

Abstract

The Cloud Radio Access Network (C-RAN) is an architecture proposition for next-generation mobile networks (5G), aimed at centralized management and processing, collaborative radio and real-time cloud computing. Such features enable this architecture to dynamically adjust the connections between Remote Radio Heads (RRHs) and Baseband Units (BBUs). However, if this feature is neglected, network problems such as blocked call and poor connection may occur. This work addresses this issue and proposes an optimized mapping model between RRH-BBU, seeking a fairer and more efficient balancing. In this sense, the Key Performance Indicator (KPI) of blocked calls was used to measure Quality of Service (QoS) metrics. In order to minimize them, an algorithm by Particle Swarm (PSO) was developed. A literature scenario composed of 19 RRHs distributed in a geographical area was used, which can be mapped in a BBU pool that manages two BBUs that have three sectors each. The initial configuration generated, on average, 80 blocked calls. Results obtained indicate a reduction of up to 100% of blocked calls and a more egalitarian load distribution among the BBUs. In addition, realistic scenarios with different user profiles were implemented, demonstrating that such factors directly impact the load generated by the BBUs and, consequently, affect their balancing. In order to verify the proposed formulation, in Network Simulator (NS-3) the same scenario used in the modeling was implemented, through the comparison of optimized and non-optimized scenarios, in order to the impact of serving more users in the network, where satisfactory results were obtained.