What is the algorithmic complexity of a getProgramAccounts()
call?
Do the Solana nodes store all the accounts ever created?
Do the nodes have to scan the entire ledger? Are there any indexes in place?
Do the length
or memcmp
filters reduce the volume of data that needs to be scanned?
Is it distributed across multiple nodes or does only the RPC node handle the query?
As the size of the ledger increases, would it be increasingly slower to do a getProgramAccounts()
call?