The idea is simple: Some big financial corporation or someone that builds an "(financial) industry standard" blockchain wants maximum trustlessness without having to rely on a set of (partly) anonymous validators (miners / witnesses), in other words they don't want to ask the world to trust the bank itself (or a number of banks) but need legal liability (someone who guarantees the finality of tx).
The solution would be to do a "subjective multisig": The bank keeps a ledger (also on a blockchain that is verifiable but with the bank or a few banks as the validators) that settles transactions if they had x confirmations on the public blockchain (say Bitcoin or Bitshares). The rule is: If there is a revision of tx history on the public blockchain that is x-1 blocks or longer or a "more serious problem" (something like a nothing at stake attack where social consensus / intervention on the public chain would be required) with the public ledger occurs the "private ledger" counts, otherwise the public ledger counts.
Remaining questions:
What are "more serious problems" and how objectively would they be detectable?
How would the "synchronization" of public and private ledger work in case there was a tx history revision that was longer than x-1?
My guess is that this is what Counterparty and Eris are doing. Just a guess no actual (insider) information.
...you could even do a several systems multisig. Not a super strong argument but Bitshares has the advantage that its consensus process is relatively unique (no other DPOS chains out there) so it would make sense to use Bitshares as one of the public chains.
...in that sense consensus is a public utility. The utility is the higher the more neutral the system is (no special interest groups that control the consensus process) and the more expensive it would be to corrupt the consensus process.
Topic: Two System (private & public) Multisig (Read 892 times)
