A block would only be accepted if it has a valid hash of the utxo set after the block's changes are applied. This would allow for you to create some sort of checkpoint at a certain block height and say, these are the block headers of all the blocks leading up to here, there is no reason for all the block headers up to this point to have been faked, as they would have been rejected (meaning there is no reason to build off of the fake chain), and the amount of computing power required to create all these block headers is greater than a bad actor could have. This is the utxo set as this block height, and I can prove beyond reasonable doubt that this is the utxo set for this block, because if you hash it it matches the utxo set hash listed in the block header.
This would make it so that after a certain block has had many blocks built off of it (say a couple years worth) there isn't any reason to keep the blocks before it.
I know that the "visa standard" of transactions per second is a bit of a naive comparison to make, but it's a goal that is easy to visualize. So lets say 1700 tps is the goal, which is what visa processes on average source (150 million per day). 1700 * 600 = 1,020,000 transactions required per block. Lets assume that an average transaction is 250 bytes. 1020000 * 250 = 255000000 or 255 megabytes per block. On average, in a day, there will be 144 blocks, meaning the blockchain will grow by 36 gigabytes a day, and 13 terabytes a year, and then 65.7 after five years. That is a very large storage requirement. I think that a good metric for how easy running a full node should be is if a typical retail store could host one. Verifying transactions is important for them when they receive so many transactions. It also helps with the retailers privacy, as they don't show their addresses to a centralized source, such as blockchain.com. Right now, point of sale equipment generally costs 1000$, plus another 1000$ a year source. After one year, that is perfectly within reach for a retailer, as 10tb hard drives are only around 350$. But after 5 years, you're getting over 2100$, and once you get above 10 years, you're getting double that, which starts to become quite out of reach (and that's just hard drive size, not taking into account other things).
I know that technology will grow to make larger and larger storage sizes cheaper, but the blockchain grows at a much higher rate than average hard drive size grows. So I think a checkpointing mechanism like this is a good solution to that.
Thoughts? I understand that
[link] [comments]
source https://www.reddit.com/r/btc/comments/neuud1/adding_a_hash_of_the_current_utxo_set_to_the/
No comments:
Post a Comment