The Internet has experienced a tremendous success. Starting from an academic (and somewhat free) communication network, it has been expanded to commercial purposes and has led to congestion. The way customers are currently charged is based on a so-called flat-rate price: they pay a fixed subscription fee to an Internet Service Provider (ISP) and have an unlimited access to the network. This simple and attractive method is nonetheless unfair since it does not discriminate users. Introducing new pricing schemes seems a valuable option for allowing congestion control and service differentiation among users or applications. While congestion hardly occur in the backbone network, we still have to investigate ways to control it in access networks, the so-called last mile problem, with a special emphasis on wireless. The challenge is therefore to design a pricing scheme representing a good trade-off between economic efficiency and engineering simplicity and that both users and providers would accept. During this talk, we will review few models for pricing bandwidth usage. We will also briefly present other contexts where pricing seems an appropriate way to incentivize users to participe by rewarding them in situations where each new user introduces an added-value to the network capability, such as for example in ad-hoc networks or peer-to-peer networks. A current research direction we will emphasize comes from the observation that there is not only a relation between customers and providers, but also a competition among providers and heterogeneous technologies, and this aspect needs to be integrated in the models and proposals. A typical example is the competition for access points at a WiFi hotspot, or the choice between different access media (WiFi, WiMax, UMTS, etc.). Similarly, pricing is also now a requirement among competitive providers themselves, which need to exchange traffic to ensure end-to-end delivery. Those points are still in their infancy and we will introduce the challenges and some proposals.