If you want to experiment with something like this you should implement it in a alt-coin. Don't try to get something like this into bitcoin before it is empirically validated.
Btw I think 50%+ mining pools is a non-issue.
The LTB Network Forum
Response to 2 phase mining. Episode 126.
Jump to bottomComments
-
Posted on July 12, 2014 at 3:27 PM
I think very large pools are most certainly non-ideal. I like the tree chains everybody-as-a-solo-miner type system best, but this seems to at the very least spread things out a great deal more.
Why do you think its a non-issue?
1 like -
Posted on July 12, 2014 at 9:00 PM
I just listened to episode 126, two phase mining; the idea that that the identities of the miners receiving the reward must be revealed and centralized by the pool is a little uncomfortable for me. Coinbase purchases are easily accessed and if bitcoin becomes illegal then there goes all our bitcoin. OF course this would never happen in the US although similarly in Australia household bank accounts were emptied of $360 million, not to mention Cypress bank accounts emptied in 2013. Now the State of Georgia US is about to do the same thing to bank accounts that have been dormant for 1 year http://www.sovereignman.com/trends/introducing-the-newest-tactic-for-governments-to-raise-cash-14566/ Naturally Georgia won't REALLY do this, right? Nor other states who are near bankruptcy like Michigan, California, etc. Well if the Government took silk road's bit coin because they were evil then what is to determine what the government decides what is evil? Obviously not using your bank account for one year is evidently very evil. The US government never took the citizens gold did they? Bitcoin needs to go global but the governments of the world are not generally concerned about their people and their assets and will take them at will. I personally feel the solution is elegant and beautiful but not applicable to reality.
2 likes -
Posted on July 12, 2014 at 11:51 PM
The bank accounts on cyprus did not get emptied. They were locked for a period of time pending a decision on a ~10% deduction as a means to bail out failing banks. (If I remember correctly, the details are getting a bit blurred).
0 likes -
Posted on July 13, 2014 at 3:25 AM
2 Phase mining is not only about restricting large mining pools but at what cost. We have to spend more processing power and electricity to achieve the result.
It would be a better if we can restrict large mining pools without utilizing extra resources.
Last Edited: July 13, 2014 at 4:08 AM0 likes -
Posted on July 13, 2014 at 5:36 AM
I'm not sure I understand this idea properly. Each of the miners would mine for a pool like today but the pool operator needs to share the private key to the address that will recieve the funds if a block is found?
Doesn't that just make every single pool vulnurable to the degree where pools are useless? After all, the pool operator would not only have to trust the miners not to run off with the money but also trust them to handle their computer security properly, since if even a single miner is compromised then the entire pools income for the next block is compromised?
Maybe the details wasn't clear enough, or is there is a glaring security issue here in that people in general are not following even the most basic of security routines?
2 likes -
Posted on July 13, 2014 at 3:14 PM
@CrisonRoze I meant the Cypress banks and Australian bank are following the same concept. My apologies for using the word "emptied" in appropriately but it was in the same context as Australian banks taking $360 million which obviously is not all the household accounts in Australia. The changes in the laws has cost me over $80,000 then another change in the laws cost me $28,000 then another change in the laws cost me $16,000 all in less than 8 years! I am at a very minimum in my level of trust toward the US government and have even less faith in any other. It appears that if you are not on welfare then you are a target to have your assets redistributed. What I am trying to bring to attention is the IF the asset is visible to the government and the financial need requires it, they will attempt to take it. Right now total obligations, entitlements, compensations and debt for the US government has exceeded 211 Trillion DOLLARS! http://www.bloomberg.com/video/80689698-kotlikoff-sees-u-s-fiscal-gap-of-211-trillion.html The government is in great financial need and everyone's assets are being reviewed.
Last Edited: July 15, 2014 at 1:46 PM0 likes -
Posted on July 14, 2014 at 12:08 AM
I've heard Andreas talk about a 51% attack not being a big issue at the moment and I tend to agree with him. My understanding is that if a pool ever got 51% and exploited it, we would all know about who did it. Furthermore they are somewhat limited in what they can do with that 51% power... they can prevent transactions from being validated and double spend their own transactions. The incentives for playing nice are much greater than what they could gain by being selfish as they would destroy their reputation and future profits in the process.
However having only a handful of pool operators pulling the strings on mining isn't a good thing in the long run and it does run counter to the decentralize everything philosophy that bitcoin started. One thing I'm still unclear on with regards to the two factor proof of work is who controls the knob that sets the ideal mining pool size? Is it baked into the protocol and hence controlled by the bitcoin core devs? Does each twiddle require a hard fork?
1 like -
Posted on July 14, 2014 at 2:19 PM
@crimsonroze I'm not sure I understand this idea properly. Each of the miners would mine for a pool like today but the pool operator needs to share the private key to the address that will recieve the funds if a block is found?
Doesn't that just make every single pool vulnurable to the degree where pools are useless? After all, the pool operator would not only have to trust the miners not to run off with the money but also trust them to handle their computer security properly, since if even a single miner is compromised then the entire pools income for the next block is compromised?
Maybe the details wasn't clear enough, or is there is a glaring security issue here in that people in general are not following even the most basic of security routines?
I think you are right about the security issues Crimsonroze. This change would require the current mining computer power to be held to a much higher level of security. So it turns the proof-of-work into a proof-of-work-trust-and-security. And not just digital security, but I'm pretty sure physical security as well, as I'm guessing the Private Key would have to be sitting unencrypted for some time in the computer.
Beyond that, I found this in a another discussion
From http://hackingdistributed.com/2014/06/18/how-to-disincentivize-large-bitcoin-mining-pools/ -Gordon Mohr "With edit - assuming the SIG() is deterministic, one SIG() try per classic hit - there seem new ways to de facto re-outsource the work. One mentioned below (Danny Hamilton) is requiring a deposit from participants, then giving each of them a unique problem (distinctive private key). If they 'steal' the reward – evident from the private key used – they forfeit the deposit, leaving the pool indifferent to the new procedure. The enforcement has changed, but the result is the same... except that the very poorest/smallest, unable to pay the deposit, are locked out of variance-reduction."
The fact that a deposit system might become the requirement does just seem like a very bad outcome.
1 like -
Posted on July 15, 2014 at 12:54 AM
With much respect to the motivation, I expect that the proposed two-phase pow would not have the intended affect. In fact, it would produce the opposite. Specifically, it will reward the lowest-latency and most-coordinated pool. Combining this with the fact that pools are being used for convenience, there also lacks an incentive to adopt the solution.
All of the largest pools do not use public keys for payout. The pool hold the funds and then, at a future time, pay out those funds in a separate. Accordingly, these pools already control the public/private keys used by the miners. So, if signed-mining (two-phase) were implemented, the large pools wouldn't change. They would just do the signing themselves. Solo miners will happily comply -- even while paying a fee to the pool.
An attack-vector is created as well. If signed-mining is required, then the fastest "signing network" will always win. This was alluded to in the original post concerning a "single large miner". A pool with 51%, and 2phase mining, will have a more significant opportunity to perform their attack. The cost of an EC signing operation is far less than the sha256 proof-of-work, thus, a 51% operator only need to invest more into their EC signing network than their competitors. Now, instead of randomness (via pow), the bottleneck is EC signatures.
Thus, two-phase will punish all who do not participate. If two-phase is implemented, the mining networks with the largest capacity to implement it will win. Bitcoin will die because these networks will be the only mining operators capable of updating their software. Why?
Almost all modern mining hardware is self-contained or runs on embedded hardware. And, most of this hardware operates on old versions of cgminer -- years old. The manufacturers write and compile the mining software specific to their miner. They do not care if it has newest mining software, they only care if it works. Right now there is about 130 petahashes of mining hardware on the market -- much of it are folks with 3-10 miners, while the rest if industrial mining. So in order for this to happen, A) manufacturers have to update software for 130PH of old miners, B) home-users have to upgrade their miners [if it's even possible; about 20% of the network runs on miners which operate on NAND and are a nightmare to update], and C) convince industrial miners to pay employees to update all their miners or install a special two-phase signing headend. For about 30% of the hardware, I would expect that mfg will not update the software -- making B & C impossible. Even if they did, what is the incentive, if pools with the fastest EC-signing networks will be the winners?
As has been mentioned, there is also the issue of a coinbase with multiple outputs and comments by Andrew Miller. There are both technical and motivational issues. Having that said, it is good to see the conversation brought up.
1 like -
Posted on July 15, 2014 at 4:08 PM
@adam I think very large pools are most certainly non-ideal. I like the tree chains everybody-as-a-solo-miner type system best, but this seems to at the very least spread things out a great deal more.
Why do you think its a non-issue?
The mining hardware is an investment in the future value of bitcoin. There is no incentives for miners to make bitcoin less valuable. And I don't think people who can get 50%+ mining power are nihilistic.
0 likes
Post Reply
Please Login to post a reply to this thread.