The Shutdown Problem: How Does a Blockchain System End?
This addresses the end-of-life management issue for blockchain developers and users, particularly for archival data, but is incremental as it builds on existing consensus theory without new empirical results.
The paper tackles the shutdown problem in blockchain systems, defining criteria for graceful termination and concluding that decentralized systems with unstable consensus like Bitcoin face significant difficulties, recommending against their use for long-lived data and suggesting hard forks to stable consensus.
We define and examine the shutdown problem for blockchain systems: how to gracefully end the system's operation at the end of its useful life. A particular focus is those blockchain systems that hold archival data of long-lived interest. We outline what it means to achieve a successful shutdown, and compare those criteria to likely end-of-life conditions in a generic blockchain system. We conclude that the decentralized nature of blockchain systems makes shutdown difficult, particularly if the system uses an unstable consensus like the Nakamoto consensus of Bitcoin. Accordingly, we recommend against using blockchain with unstable consensus for any data whose value is likely to persist beyond the life of the blockchain system. For any such systems that are already in operation, we recommend considering a hard fork to implement stable consensus. Such consideration needs to happen well in advance of the system's end of life.