Enhancing throughput in centralized radio access network by intelligent controller placement

In the modern society, Internet plays a key role in everyday life and, to sustain its exponential traffic growth and Quality of Service (QoS) requirements, network operators evolve the mobile and fixed network infrastructure continuously. Specifically, mobile devices have been very prosperous lately and network operators are migrating from 3G to 4G networks, which aim at improving network performance and reducing cost, moving a step forward to the new generation network (5G). Among the main trends for 5G, we focus on two concepts. The first aspect is BaseBand Unit (BBU) hotelling often associated with the Centralized Radio Access Network (CRAN), which consists of splitting physical layer (i.e., L1) functions from L2/L3 operations performed on radio signals. In BBU hotelling, the BBUs are separated from Remote Radio Heads (RRHs) and are centralized into a common location, called hotel. As a consequence, the infrastructure is reduced, though a new traffic type (i.e., the so-called fronthaul), is introduced between BBUs and RRHs, and it must be carefully managed due to its stringent requirements in terms of latency and bandwidth. The second focus of this work is on Coordinated Multipoint (CoMP), which is a technology used to coordinate clusters of cells from a Radio Coordination Controller (RCC), aiming to enhance throughput and reduce interference for users at the cell edge. The X2 cell interface manages the coordination traffic that flows between cell sites within a cluster and the corresponding RCC. The throughput gain given by the cluster coordination is affected by many factors, e.g., cluster size, X2 latency and coordination technique. The objective of this thesis work is to prove that an optimized RCC placement and clustering can have better results in terms of coordination gain and interference reduction with respect to a simple RCC placement. To reach our goal, we implement a Clustering, Routing and RCC placement optimization problem for an access/aggregation mobile network. The optimization problem is written with Interger Linear Programming (ILP) method and takes part to a complex solving procedure, as it presents scalability issues. In our Clustering, Routing and RCC placement problem, we tested several dimensions: CoMP coordinations techniques (i.e., Coordinated Scheduling, Joint Transmission), scenarios (i.e., intra-site, inter-site), RCC processing latency (i.e., fixed, variable) and geotypes (i.e., urban, suburban, rural). In detail, we compare the coordination gain in different geotypes and scenarios with and without BBU hotelling. We prove that, in some specific cases, an optimized RCC placement and an optimized cell sites clustering can bring further benefits with respect to the RCC placement at the core CO.