Topic: New database for blockchain - page 2. (Read 456 times)
Or to ask it in a slightly different way... what do you believe you will gain from this setup as opposed to the current setup?
That's extremely inefficient. You will end up with a lot of files that are so tiny (176-300 bytes) and opening each file and reading it is still consuming the biggest percentage of the total time.
My idea is through the file name of the blockhash is to find and open the blocks very quickly.
The current way of storing multiple blocks in a file and storing an index for their location is fast enough. Why do you need more speed? 
Do you really think that your database is better? There is more to a database than how to organize the files.
I want to create a new database for the blockchain.
I would like to create a file for each block.
eg: 000000000019D6689C085AE165831E934F763AE46A2A6C172B3F1B60A8CE26F.Block
That would be more than 700,000 files.
NTFS can be 4.294.967.295 files in a folder.
So theoretically it should go.
Question:
Has anyone ever tested such a thing?
Would the file system crashed or something similar?
My idea is through the file name of the blockhash is to find and open the blocks very quickly.
Ideas and suggestions are welcome :-)
I would like to create a file for each block.
eg: 000000000019D6689C085AE165831E934F763AE46A2A6C172B3F1B60A8CE26F.Block
That would be more than 700,000 files.
NTFS can be 4.294.967.295 files in a folder.
So theoretically it should go.
Question:
Has anyone ever tested such a thing?
Would the file system crashed or something similar?
My idea is through the file name of the blockhash is to find and open the blocks very quickly.
Ideas and suggestions are welcome :-)
Jump to: