Post
Topic
Board Development & Technical Discussion
Re: 'Blockchain chunking' - can that be done?
by
JPage
on 16/12/2016, 20:51:43 UTC
Special Archive Nodes.  you know if the archive has been altered, because the hash is passed into the new chain.  Any change to the old chain would render that verification hash invalid.  If there are multiple versions of the archive, only the correct one will produce the right hash.  

How is a "Special Archive Node" any different than a regular node today?  Since the integrity of the network depends on the existence of these "Special Archive Nodes", won't it be just as important to have them as it is to have regular nodes today?

Doesn't that mean that we effectively already have this?  Aren't you essentially just renaming "full nodes" as "Special Archive Nodes" and renaming pruning nodes as "Full Nodes"?

I don't know about the function of 'pruning nodes'.  Maybe we do already effectively have this. Smiley  I admit I haven't looked at it very closely.  But I imagine that somehow carrying around microtransactions for all of time is a bad idea when we finally get to 7K/transactions per second.  There has to be a mechanism to throw the old record away and start with a current balance sheet.  Just wondering.

Nodes running Bitcoin Core (or any up to date variant) can set a "Prune" size, and they will only store recent blocks up to the indicated size they are willing to hold.

Meanwhile, Full Nodes (nodes that don't turn on Pruning) store the entire historical blockchain so that new nodes can synchronize and get caught up.

Satoshi already pointed out in the white paper that (as long as there are some Full Nodes that store the entire history), most clients can prune from their blockchain all transactions that have had all their outputs spent.  All that is needed for a full node to verify new transactions and blocks is the UTXO and the transactions that supply the UTXO.  The rest is only needed if you want to be able to look at old transactions or if you want to help new nodes get started.



Then why does anyone give a damn about blockchain bloat?  It seems like that system would still work fine even when the blockchain is 50,000TB.  I am sure we'd still have 10 full nodes even with a HUGE blockchain.