On Matching Users to Specialised MNOs in Service Tailored Networks of the Future

In this paper, we investigate a network model in which entities called subscription brokers group service level agreements with multiple specialised mobile network operators (SMNOs) into a single subscription bundle, with a fixed data allowance that can be used by the subscriber as needed across any of the SMNOs included in the bundle. Each SMNO operates a network that is designed to meet the demands of a particular service area or vertical industry. We demonstrate the performance benefits of such a model, allowing users to choose SMNOs according to the needs of the service that they are using. In particular, we focus on how to perform the matching between users and SMNOs in a bundle, adopting the Gale-Shapley matching algorithm. We argue that a stable matching is needed to ensure that both SMNOs and users are incentivised to adopt the broker-based model. We outline a framework based on the concept of utility for devising the preference lists of users, while the approach we propose for building the preference lists of SMNOs can differentiate between different classes of users based on the price they pay for their subscription. We evaluate the performance cost in terms of utility of achieving stability compared to a sum utility maximisation matching approach, showing that this cost is largely borne by the lower priority users. Overall, the proposed broker-based model performs at least as well as any one SMNO for lower priority users, and outperforms any one SMNO for higher priority users.