So I've been thinking about this scenario, you are a survivor of a global disaster. Years later you find a running solar powered bitcoin node and want to restart the chain. You have a few running computers that you found and starlink is still up but not 24/7. You also built a local wifi wisp and there are others on the network that wish to run nodes.

What would the bare minimum infrastructure need be to re-start the chain? What would you need to do to get a mining computer to sync and start mining as there is no setgenerate available in the cli?

I will do a write up on the procedure if it's possible.

1 Answer 1


You'd have at least the following different issues:

  • Finding the best chain: When you investigate the blockstate database of the computer, you'll find out when it was disconnected from the network, but you have no way of determining whether there is a better chain tip somewhere. If networks eventually reconnect, only the chain with the most work would prevail, and especially all transactions based on block rewards in your network might get invalidated. If the Bitcoin satellite is still operating, you could learn about the best chain by putting up a satellite receiver. The blockchain has also been broadcast per radio frequencies before.
  • Submitting transactions: Even if you hear about new blocks, e.g. via satellite or radio, you might still not have a way to broadcasting transactions to the network. Since transactions are fairly small, you could transfer them even without a global internet connection, e.g. via ham radio, printed out, or via USB-stick carrying pigeons, if there is other regions that still have further reaching network connections.
  • High Difficulty: If the network is not just split, but you're actually all that's left of the Bitcoin network, the difficulty for mining is likely way out of reach for what you can achieve with just a few computers. Mining today is done with specialized hardware which is magnitudes more efficient than general computing devices. You may need to amend the rules to lower the difficulty so your mesh network can start to find blocks again. Since your new rules will be incompatible with any existing Bitcoin installations, you may want to do a clean hardfork that prevents replay attacks, switches the proof-of-work algorithm and either starts from a clean-slate or airdrops the current Bitcoin balances to the same addresses. Assuming that many little pocket societies do the same, it would be hard to return to a shared standard if and when groups reconnect.

So, if you're just starting your own new network locally, you only have to maintain some connectivity in the network and appropriately lower the difficulty. If all the connections are intermittent, you might want to increase the block intervals in your new coin to lower the number of chainforks that need to get resolved. If you're trying to rekindle the actual Bitcoin network, you will need to gain access to substantial SHA256 hashing power to reach the next difficulty adjustment, and you'd want to connect as far as possible to find a better chain as soon as possible to not waste too much resources.

You may also find these related topics interesting:

  • Thank you for the detailed answer Murch! So as I see it, on a small scale the BTC network would be doomed. Even if you had say a large maybe hydro powered mining operation it might take years to find the next block.
    – Ron
    Aug 26, 2022 at 16:33
  • Yes, both sourcing sufficient energy and hashrate would be problematic. It would also not only affect the next block. The difficulty only adjusts every 2016 blocks, and only by 0.25–4×. It would probably take a number of difficulty periods for the difficulty to come down to levels where blocks would be found every ten minutes, unless the difficulty were adjusted with a hardfork.
    – Murch
    Aug 26, 2022 at 18:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.