Differential Pricing of 5G Network Slices for Heterogeneous Customers

5G wireless networks promise to enable new kinds of cellular use cases, by offering different network slices to users with different needs. While pricing network bandwidth is relatively straightforward when all users care about data rate, more sophisticated pricing strategies are likely to emerge when some of the customers acquire a network slice with the desire to optimize for a different metric. We consider how a 5G wireless provider may set per-class differential prices to maximize its profit when offering different network slices to different classes of customers. We formulate the problem using fundamental economic principles and present a Drift-Plus-Penalty algorithm to solve the problem in a dynamic setting. We show, through simulation and analysis for a two-class network with latency and throughput-oriented customers, that some surprising phenomena may occur under certain conditions, such as an overall reduction of total resources sold when the relative number of latency-oriented customers grows.