Optimal Dynamic Fees for Blockchain Resources

Users of public permissionless blockchains can modify the shared state of the network through transactions that are executed by a set of nodes with limited computational resources. To allocate resources among competing transactions most blockchains use transaction fees. Initial transaction fee mechanisms in the Bitcoin and Ethereum blockchains relied on users bidding for transaction inclusion as the main way of pricing congestion. Moreover, all computational resources were bundled into a unique virtual resource ("gas") with fixed relative prices hardcoded in the protocol. Current R&D efforts are focused on improving transaction fee markets along two directions: (1) setting a minimum dynamic base fee (henceforth also called price) that is adjusted by the protocol as function of user demand and (2) unbundling resources so that different resources can be individually priced and their relative prices can also efficiently adjust with demand. In this paper, we propose a new framework for choosing a resource pricing policy that makes significant progress across both directions. We consider the practical problem of a blockchain protocol that has to jointly update the prices of multiple resources at every block. We assume that the type of resources being metered and priced, as well as the block limits and sustainable targets for each resource, are pre-determined.