The context is here

I don't know how exactly the new blockchain candidate was calculated in Bitcoin Network, so I guess it was based on a Merkle candidate of (a) bitcoin client/s. Also I presume that is by referring the diff of network, a client hashing the/a (golden) nonce, and then propagate it to the network, if lucky enough other nodes think this nonce is ok (the first), somewhere magic happened you get the coin (please enlighten me about this magic).

Please correct my understanding.

So (if you know Hadoop)

Despite how actually SHA(Midstate(nonce++)) was implemented in a Reduce procedure (hashing/sec/energy), how could I generate new Merkle tree(s) and midstate candidate(s) in Hadoop way (huge number of them simultaneously)?

Thank you for your reading (hard to understand).

Problem (if prior has no big problem)

What I am proposing here is, if all this works, how could a mining pool like this benefit from Hadoop? (if it could)

(If it really works, I'd implement raspberry pi to hook up hundreds of SHA asis of my own in distributed fashion.)

  • If it really works, i'd implement raspberry pi Don't. It's less efficient than an Intel CPU. – Nick ODell Jul 27 '13 at 14:58

Map Reduce isn't the best idea, because distributed mining is solved and works well. The magic of mining is described here, and when that is complete that Bitcoin protocol allows for 25 new Bitcoins to be send to any address. This is encapsulated into a block that is either accepted (and used as the "previous" block), or rejected, for example if the protocol rules are broken.

Your time is best spent running your own mining pool.

  • What if the mining midstate pragmatically is not the best shot, and some pre-cal of blockchain could be done in MapReduce? while the diff is increasing could targeting the nonce from reverse space could be possible to implement? – einzeln00 Jul 27 '13 at 13:58
  • 1
    targeting the nonce from reverse space Huh? – Nick ODell Jul 27 '13 at 14:57
  • @einzeln00 You need to learn more about how mining works, and how it relates to the blockchain. No pre-calculations are possible in mining, the merkle tree, the nonce, or the midstate. Difficulty isn't "real" it's really the "target" and that can't be optimized or distributed. Read the link I posted, and if not my answer read some other answers on that page – halfbit Jul 27 '13 at 15:26

Not the answer you're looking for? Browse other questions tagged or ask your own question.