Most data storage systems have a way to automatically decide the physical location that data will be stored. (That's why, for example, you don't have to calculate something about sectors and tracks when you save a file on your computer.)

If you do automatic allocation in decentralised storage, then you also need to automate the system that distributes payments to the service providers — Filecoin-style manual peer to peer negotiation is not possible. What is needed is a mechanism in the sense of microeconomic theory. This mechanism needs to have good properties if it is to foster a healthy service provider population and deliver good service and efficient prices to users.

I'll give an introduction to this subject in the context of the EthSwarm Protocol. In EthSwarm, as well as providing a fixed amount of capacity, service providers must buy revenue shares from a contract at a fixed price in a kind of Tullock contest. Perhaps surprisingly, even though this allocation system has nothing to do with the amount or quality of service provided, under good conditions it results in revenue share flowing to more efficient operators, resulting in more competitive pricing.

Target Audience: people researching and building decentralised services