A matching game for data trading in operator-supervised user-provided networks

In this paper, we consider a recent cellular network connection paradigm, known as a user-provided network (UPN), where users share connectivity and act as an access point for other users. To incentivize user participation in this network, we allow the users to trade their data plan and obtain a profit by selling and buying leftover data capacities (caps) from each other. We formulate the data trading association between buyers and sellers as a matching game. In this game, buyers and sellers rank each other based on preference functions that capture the buyers' demand for data and QoS requirements, the data available for purchase from the sellers and energy resources. We show that these preferences are interdependent and influenced by existing network-wide matching. For this reason, the game can be classified as a one-to-many matching game with externalities. To solve the game, we propose a distributed algorithm that combines notions from matching theory and market equilibrium. The algorithm enables the players to self-organize into a stable matching and ensures dynamic adaptation of price to data demand and supply. The properties of the resulting matching are discussed. We also calculate operator gains and the benchmark price that will encourage users to join the UPN. Simulation results show that the proposed algorithm yields average utility per user improvements of up to 25% and 50% relative to random matching and worst case utility, respectively.