Cost Effective Provisioning of 5G Transport Networks: Architectures and Modelling

The next generation of mobile network (5G) has to face a completely new set of requirements coming from novel services. Massive machine type communications, enhanced mobile broadband, ultra reliable low latency communication will be supported by single infrastructure. Mobile network operators are in need of a flexible network capable of supporting services with a wide set of different requirements over the same physical resources, possibly at the same or at a lower cost than today. Centralized radio access network (C-RAN) architecture is a promising solution to improve both network flexibility and scalability. In C-RAN, baseband processing units (BBUs) are decoupled from remote radio units (RRUs) at the antenna sites and are placed in one of few selected locations, called BBU hotels. Thanks to the centralization, more efficient hardware can be employed, advanced radio interference management techniques can be implemented, cooling and power supply units can be shared, and network maintenance is simplified. However, the centralization of BBUs requires high capacity and low latency dedicated links to transport data, known as fronthaul links. This may be expensive and calls for novel deployment strategies to contain the costs. This Ph.D. thesis investigates the cost-efficient and resilient design of C-RAN. Minimization of network equipment as well as reuse of already deployed infrastructure, either based on fiber or copper cables, is investigated and shown to be effective to reduce the overall cost. Moreover, the introduction of wireless devices (e.g., based on free space optic) in fronthaul links is included in the proposed deployment strategies and shown to significantly lower capital expenditure. The adoption of Ethernet-based fronthaul and the introduction of hybrid switches is pursued to further decrease network cost by increasing optical resources usage. Finally, the problem of single BBU hotel failure is addressed and included in the optimal deployment of BBU resources.