That's actually very close to what we are trying to solve in competition. One thing that made us choose a blockchain-based approach instead of simply keeping another copy in something like S3 is that we wanted the integrity reference itself to be outside the organization's normal infrastructure. The company could operate one of the nodes, but the verification would still rely on the other independent nodes in the network rather than on a storage system controlled by the same organization. We also keep the integrity state time-based, so instead of only knowing that something in the historical data changed, we can go back to a specific time window and verify that period against its previously recorded state. I honestly hadn't considered the S3 + Object Lock approach before, so I'm interested in how you would compare that with an externally anchored Merkle root. Our project is already built around this idea and I'd really like to show you the approach rather than try to explain all of it in the thread.
... View more