A Mechanism for Price Differentiation and Slicing in Wireless Networks.

Slicing has been introduced in 5G networks in order to deliver the higher degree of flexility and scalability required by future services. Slice tenants such as virtual wireless operators, service providers or smart-city services will be able to book a share of the infrastructure, possibly including storage, computing capacity and link bandwidth. However, 5G slicing is attractive for infrastructure providers as long as they are able to generate revenues, while at once satisfying the tenants' competing and variable demands and coping with resources availability. This work proposes a flexible mechanism based on a multibidding scheme for 5G slice allocation. It is able to attain desirable fairness and efficiency figures in order to serve slice tenants and associated mobile users. Built on the notion of normalised Nash equilibrium, it is also provably overbookingfree even though the players' bids are oblivious to infrastructure resources constraints. Also, it is compatible with standard radio access schedulers used in modern mobile networks. Finally, a practical algorithm is proposed to drive the system to the socially-optimal operating point via an online procedure rooted on a primal-dual distributed algorithm. Numerical simulations confirm the viability of the mechanism in terms of efficiency and fairness.