Hacker News Reader

Comment by kysar_tom | Hacker News Reader

You could, but a relational database isn't distributed upon 5,000+ worldwide nodes in an immutable data ledger. That's the value the Bitcoin blockchain provides, it allows you to stick data into a transaction that will be there forever. The Bitcoin blockchain provides an economic incentive to keep the data accurate and accessible, as it would be (near) impossible to pull off a 51% attack in todays day and age.

Many databases can be modified by the ones who centrally operate it. The Bitcoin blockchain is the worlds first distributed global database that takes the burden of trust off your internal systems and records.

View on HN

No; using the Bitcoin blockchain to store medical records is a terrible idea.

Yes, they are robustly distributed on each of 5,000 worldwide nodes.

But, think about that! How's that going to work for 10 million patients' MRI records? Each MRI being a series of high res images?

That's not possible with Bitcoin - its not designed to store near that quantity of data - its not even desirable.

I don't know what the advantage of Blockchain is with healthcare, but its certainly not storing your medical records on the Bitcoin blockchain. [Standard disclaimers about predicting the future apply.]

Maybe a hash of them, so you know they aren't tampered with, or that sort of thing - but that seems like a second order problem.

No one is advocating putting medical records on the blockchain. That's crazy.

He's talking about using a technology like Tierion or Chainpoint to anchor a hash of the data in the blockchain. This can be used to verify the integrity and approximate timestamp of the data.

http://tierion.com http://chainpoint.org

> No one is advocating putting medical records on the blockchain. That's crazy.

The post I responded to seemed to be.

It implied the data would be "distributed upon 5,000+ worldwide nodes in an immutable data ledger", and said it would:

"be there forever" "keep the data accurate and accessible"

I think its legit to call this out. Blockchain is a great technology, but its being hyped as if its the world's only distributed database.

> This can be used to verify the integrity and approximate timestamp of the data.

That may indeed be beneficial, but isn't the panacea that is being touted.

"That may indeed be beneficial, but isn't the panacea that is being touted."

We agree. That's the primary conclusion of the report.

Should have been more clear. No records should be stored on the Bitcoin blockchain, only a hash of the data. Using a service like Tierion, a hash of your data is anchored into the Bitcoin blockchain, and Tierion gives you back a blockchain receipt. That blockchain receipt is a cryptographic proof that can be used to verify the integrity of that data at a later point in time.

I still don't see the point in a 'permissioned blockchain' without proof of work. If you just put Paxos / Raft / any other consensus protocol on top of a 'blockchain', all you're basically doing is reinventing distributed databases.

Tierion uses the Bitcoin blockchain for these reasons. Proof of work is what really secures your data.

Is this a trustless system? If so, what is the game theoretic aspect of it, what nodes do the mining and what are their incentives? If it is not, what prevents you from implementing a traditional distributed system with traditional consensus protocols? I'd wager that traditional methods are much more performant than the BitCoin blockchain in that scenario.

> You could, but a relational database isn't distributed upon 5,000+ worldwide nodes in an immutable data ledger.

The DNS would like a word.

Reply on news.ycombinator.com