Thank you, for finally putting this out. I was very confused as to why am I the only one having these strong feelings against how the tech works.
I think the guys over at bigchaindb present quite an interesting solution to the problem Blockchain is trying to solve, without giving up scalability and speed.
I'm happy to discuss in detail if there's interest, but essentially the misconception is that all previous states need to be kept. Ethereum uses the blockchain as a state machine where each block is a state transition and the current view of state (world state) is kept as a piece of data connected with an account.
Being unable to audit Ethereum's state at any historical block is absolutely not necessary. You only need the most recent state for value transfer or full validation of new blocks.
You certainly can't audit the historical state with any Bitcoin-based blockchain, which don't even have a root commitment for the most recent state.
If Etherum generates 51GB every three years at 24,000 TPH then Visa isn't going to produce significantly less than 13TB every 3 years at 6 million TPH. Lots of transactions require lots of storage. Visa doesn't pull any magic tricks, it just stores more data. If storing more data isn't viable for blockchains then they will have a hard time against centralized databases.
Hi, I'm not sure where you get that "All previous states do need to be kept if you aim to audit the chain..." but you are wrong.
We only need to keep the blocks with the transactions which are replayed every time you sync a new node and all (edit: intermediate) state transitions are re-created from the genesis file until today.
For most cases, the cryptographic guarantees you can get synching only the block headers as a SPV node are strong enough.
Think of it this way, let's say you are obligated to carry 1 valid photo ID with you. Would you carry only the latest and valid version or would you carry every single issue of, let's say, your passport?
Speaking of perspective, Visa exists since the late 50s and Ethereum for about 5 years. The next iteration of the network will be focused on scalability and privacy technologies and we will surpass the 1000 tps on-chain and have the off-chain capability for millions of tps.
This is why I laughed at Bitcoin Cash's proposed "We'll make our blocks as big as demand is", because long term that doesn't scale if every dedicated host needs the full ledger.
8 comments
[ 3.3 ms ] story [ 35.0 ms ] threadI think the guys over at bigchaindb present quite an interesting solution to the problem Blockchain is trying to solve, without giving up scalability and speed.
https://youtu.be/1NHHmHVCWy0
No the blockchain size hasn't reached 1TB yet and doesn't seem that close at ~51GB at the moment. http://didtheethereumblockchainreach1tbyet.5chdn.co/
I'm happy to discuss in detail if there's interest, but essentially the misconception is that all previous states need to be kept. Ethereum uses the blockchain as a state machine where each block is a state transition and the current view of state (world state) is kept as a piece of data connected with an account.
Just the chain itself already being at 51gb is also a problem since it’s so new and has relatively non-existent volume.
Ethereum is pulling 24,000 transactions per hour on average.
Compare that to say Visa which is doing 6,000,000 per hour.
You certainly can't audit the historical state with any Bitcoin-based blockchain, which don't even have a root commitment for the most recent state.
We only need to keep the blocks with the transactions which are replayed every time you sync a new node and all (edit: intermediate) state transitions are re-created from the genesis file until today.
For most cases, the cryptographic guarantees you can get synching only the block headers as a SPV node are strong enough.
Think of it this way, let's say you are obligated to carry 1 valid photo ID with you. Would you carry only the latest and valid version or would you carry every single issue of, let's say, your passport?
Speaking of perspective, Visa exists since the late 50s and Ethereum for about 5 years. The next iteration of the network will be focused on scalability and privacy technologies and we will surpass the 1000 tps on-chain and have the off-chain capability for millions of tps.