Author

Topic: "Tarball" of blocks to speedup first full sync (Read 846 times)

newbie
Activity: 28
Merit: 0
August 11, 2015, 10:12:48 PM
#6
We used to; but since 0.10 the blockchain is now download in parallel and verified concurrently. Using the seperate download, even via bittorrent, then loading it is now usually slower than just syncing directly.

Thanks for confirming this, Glen
staff
Activity: 4242
Merit: 8672
We used to; but since 0.10 the blockchain is now download in parallel and verified concurrently. Using the seperate download, even via bittorrent, then loading it is now usually slower than just syncing directly.
legendary
Activity: 1027
Merit: 1005
Awhile back I offered to sell a USB stick with the updated blockchain on it but I couldnt find a way to prove that it wasnt infected with virus or some wallet stealer.

Here is the blockchain downloadable via torrent: https://bitcointalksearch.org/topic/ann-bitcoin-blockchain-data-torrent-145386

EDIT: after looking at the thread again it seems this is actually slower than letting the client sync... sorry
newbie
Activity: 28
Merit: 0
I agree about the indexing effort being unaffected, and that blocks are not terribly compressible. But fetching blocks in chunks of 500 blocks is quite a bit of network overhead which could be positively impacted by this scheme.

Of course if someone had some data on how much of total wall clock time is for different tasks that would help...
staff
Activity: 3458
Merit: 6793
Just writing some code
Because the data is still there and the same size. To download all of those blocks in those chunks is downloading the same amount of data from the p2p network. It won't be any faster and it all really depends on your network speed. The time for indexing all of those blocks is also the same because that is based on the cpu. Also, I remember someone mentioning that compression does not work well with Bitcoin blocks since there isn't much data that can be losslessly compressed.
newbie
Activity: 28
Merit: 0
 I recently download Bitcoin Core, and been waiting for it to sync up... it is a highly frustrating experience to wait for so long. It made me wonder ...

If the genesis block could be baked into the code, why can't the next 300,000 blocks be made available in 50K chunks? It's not like they are going to change or anything...

Trust is not an issue as the blocks will still be validated and accepted as today. It will cut out the network handshakes and improve startup times. People can even tshare the blockchain on thumb drives in local meets.

Yes, it is not a 'pure' and 'elegant' bitcoin experience.. but do you think it will make a significant impact on initial sync times?
Jump to: