A Stackelberg Queuing Model and Analysis for the Emerging Connection-based Pricing in IoT Markets

As the demand for low-power wide-area (LPWA) communication increases in IoT market, LPWA communication services with a variety of pricing schemes have been proposed by network operators worldwide. Among these pricing schemes, connection-based pricing (CBP) differs itself from the rest data-volume oriented schemes by charging LPWA applications by the connection requests devices make rather than the amount of traffic they use. Intuitively this novel pricing scheme could help network operators monetize massive connections better and alleviate the accompanied network burdens on the control plane. However, its implications on the network operator have not been fully studied to date. Existing models on data volume based pricing schemes either miss the stochastic dynamics or focus on players with elastic utility functions, making them not applicable to LPWA cases anymore. In this paper, we first build a new Stackelberg queuing game model to better describe the connection-based pricing. Our analysis on this model then reveals several insights that understand CBP in a game-theoretic queuing environment. Our conclusions include: 1) CBP is more appropriate for loss-sensitive applications, small-sized network operators, or network operators in the early stage of the LPWA market. 2) CBP could help reduce congestion when compared with the traditional flat-rate pricing method. 3) CBP is a form of congestion-based pricing scheme, where the congestion level has to be incorporated in pricing to guarantee profit maximization.