Radio Access Network and Spectrum Sharing in Future Mobile Networks

Abstract

Future mobile networks will be characterized by wide-spread resource sharing, and a more nuanced view of what it means to offer a mobile service. The traditional view of a mobile network as a system that simply delivers voice and data to a geographical area no longer reflects the reality, as major changes have occurred to the structure of ownership and control in mobile networks. Contemporary mobile operators share their networks with other operators and allow other entities such as mobile virtual network operators to utilize their resources under their own brand to their own subscribers. What is more, a new breed of a virtual operator has evolved -- one that is an over-the-top service provider, an infrastructure provider and a mobile virtual network operator at the same time. All these trends give rise to a plethora of resource management problems, some of which we address in this thesis. Specifically, our key research questions are: What statistical models are representative to inter-operator resource sharing in spatially distributed mobile networks? What is the efficiency of radio access network and spectrum sharing in spatially distributed mobile networks? How does network resource sharing affect utilities obtained by traditional mobile network operators and over-the-top service providers? To address these questions we apply a range of analytical tools, including stochastic geometry, optimization and game theory, and, whenever possible, we populate our models with real mobile network data.

In consequence, we find that there is a high degree of spatial dependence between the infrastructure deployment made by multiple operators. This dependence, which includes both spatial clustering and repulsion, occurs over and over again for similar areas in different countries, and can be best modelled using clustered point processes, such the log-Gaussian Cox process. When resource sharing is applied to spatially distributed mobile networks, we see that this spatial dependency has a significant impact on the efficiency of resource sharing. Furthermore, we utilize stochastic geometry to assess the trade-off between spectrum and radio access network sharing. What we observe is that each type of sharing has its own distinctive trait. While spectrum sharing alone provides an increase in the data rate at the expense of coverage, infrastructure sharing results in improved coverage. When the two sharing types are combined, we observe significant increase in the average user data rate and minor improvements to the network coverage as compared to the non-sharing scenario. Consequently, combination of the two approaches does not linearly scale the respective coverage and data rate gains. Yet, as we show, even a simple mechanism of coordination in spectrum sharing allows operators to reduce the impact of this trade-off and improve their coverage without hampering the user data rates. Additionally, we observe that the way in which spectrum is being shared makes a difference to both the coverage and data rate performance. In particular, when coverage is the main concern, it is more beneficial for operators to allow for their transmitters to dynamically select among the pooled spectrum bands, rather than simply bond the available channels.

When decisions by the entities involved in the current communications market are considered, we find that considering a multi-operator network infrastructure as a collective may lead to building up coverage provisioning networks which require just a fraction of all the available transmitters, with low density areas benefiting considerably from infrastructure sharing, which may make the case for infrastructure deployments in currently under-served rural areas. Furthermore, when new network deployments are to be made, we show that it is reasonable for over-the-top service providers and mobile network operators to cooperate. This cooperation will be primarily conditioned on the price of the radio access infrastructure and, to some extent, also on the spatial clustering in mobile demand.

All of this builds groundwork for the vision of future wireless mobile networks, which we term Networks without Borders. This vision is based on a marketplace of virtual network operators that construct networks from a pool of shared resources, such as base stations, spectrum, backhaul, and core network servers. These resources are sourced from traditional industry players as well as from individuals. As we show, the proposed vision results in a substantially different mobile communications value-chain, which results also in a number of socio-political challenges.