distributed agreement on data #3
Description
agreement on data is a crucial topic.
since there's no enforcement on the main chain, as proposed by Peter Todd in delayed-utxo-commitments, we must do something else.
remembering this is a substitution protocol on indexes services, and we must rely on some authorities, we can anyway split the responsibilities across multiple entities (and adding optional trust networks).
starting from the basic assumptions that:
mempoolcowill be the first authority to release the utxo data & patches.- anyone can verify the data & patches independently, running the full utxo indexing mode.
- transparent well-reviewed consortia could be a good source of trust, even if not math enforced (reputation).
it comes that:
- nodes have to be bootstrap with a key to recognize authorized signers of the utxo set (trustnet)
- nodes must validate the consistency and agreement of data, by using the trustnet keys, which must be always 100% by design
- the more the trustnet is extended, the best it is, cause peers will independently verify that all the authorities agree on the new data by simply downloading signatures
- the more is extended, the more is resilient, cause different authorities can cooperate to do the same job, so if (when) at some point mempoolco goes offline for a while, other authorities will still release new data as they were already doing.
- a new token could be issued to do fancy things (ok, kidding, but it seemed we were going that way at some point lol)
- clients could add their own authorities like their own, or friendly trusted full nodes signing the utxoset, or electrum indexing services which decide to serve this protocol as well