Bitcoin War: The First Real Threat to Bitcoin?
For most Bitcoiners, it is a well known fact that there is a significant risk in the decentralized peer-to-peer currency pertaining to hashing power. In order to maintain a perfectly democratic internet currency, no one single entity should ever have control of 51%, or greater, of the total network hashing power.
Today, one of our researchers discovered that according to Blockchain.INFO, a miner at 85.214.124.168 currently has approximately 15% of the total hashing power. This, in itself, is every day news. However, the strange or even frightening fact is that it is producing empty blocks (single transaction blocks). If this 15% turned into 51%, it could have the potential to kill bitcoin. Why are they doing this? There are a few possible reasons:
1. The entity may have discovered a method for increasing mining ROI, and essentially, is earning its 50 bitcoins per block much more quickly than others. In general when finding a block, hashes for every transaction must be computed. When computing 1 transaction per block versus 100, you can imagine the latter would be more costly than the former. However, this means that the entity would not receive any fees for processing transactions. It is difficult, at the current time, to determine whether this would be beneficial.
2. The entity is willing to blow money on mining these empty blocks. Essentially, this could lead to a complete stop in bitcoin transaction processing. If the entity obtains 51% of hashing power and fully stops processing transactions while mining against only its own blocks, the block chain will become useless. Some people who might do this include governments, banks, competing currencies, or ridiculously wealthy and bored individuals who have a vendetta against bitcoin.
3. This could also be a botnet that does not wish to deal with the hassle of constantly sending all of the current transaction information to its zombies. This would be more for coding simplicity rather than for financial gain.
As of today, there is still very little risk. Additionally, assuming this entity falls under the #1 listed above (i.e., not entirely malicious), the worst thing that will happen is that bitcoin transaction confirmations will be slowed down by whatever percent of hashing power they are “contributing,” and Jerry McGuire will yell “SHOW ME THE MONEY!”
So what is it? Is this entity generously increasing their ROI, or is it attacking and taking over? With the recent security advisories and, of course, the widely publicized hacks, it looks like the WWW (Wild Wild West) is in full effect.
Comments are closed.
69 Comments
85.214.124.168 is located somewhere in Berlin, Germany. Are you sure it isn’t Deepbit?
After all it turned out to be Eligius Mining Pool :-)
Bitcoin doesn’t break if someone with 51% of the hash power doesn’t include transactions in their blocks, unless they try to fork the chain (only mine on top of their own blocks), which is not what this guy is doing.
The other 49% can still include transactions, so your transactions will just take an extra 10 minutes on average for the *first* confirmation.
A deposit to MtGox that normally took 60 minutes on average, would now take 70 minutes on average.
yep, right. Bitcoin transactions are not completely dependable on mining new bitcoins. That would be a flaw by design :-)
Well at first, yes I thought that could be a problem, but now I don’t think so any more:
Well lets assume this strange miner gets 60%. Well ok, then only 40% of the blocks contain transactions and only these 40% get transaction fees. The strange miner will either realize some time that getting transaction fees is more profitable, or it will become too unprofitable to continue when the BTC reward cuts in half.
The problem that was clear from the beginning of BTC is that if someone gets over 50% of the hashing power and this miner makes invalid transactions. That would, I suspect, be followed by a short fractioning of the blockchain until the majority of miners would reject all blocks with invalid transactions and mine a different chain. Then all the money someone put into hashing invalid blocks would be wasted. That alone will prevent such an attack.
The most important thing about this in my opinion is that it illustrates how easy it is to monopolize the bitcoin network.
If you can get 15%, well 51% is only about 3.4 times as hard.
The question should be not if a 51% monopoly occur, but what will the monopolist decide to do once he accumulates 51%.
This could happen quite suddenly. Suppose for example that the btc price crashes again and GPU mining becomes unprofitable even with cheap electricity. Most miners will exit if they can’t recoup electricity costs. Most mining is still overwhelmingly GPU mining. If this guy is a botnet, electricity is irrelevant to him. If he is a botnet, a price crash could send him directly to 51%. If investors think that 51% control leads to the destruction of bitcoin (I don’t but many do), then a price crash would be a self-fulfilling prophecy.
Under a self-fulfilling prophecy, the bitcoin price has multiple equilibria. Price can be high with a diffused hashing control equilibrium and low with concentrated hashing control equilibrium. As a speculator, you can earn big bucks by investing money to switch between equilibria. All you need to do to profit is put a large chunk of money in shorts. Others will worry that heavy shorts could get the price drop and mining exit rolling, giving mystery miner more control. If you believe that most of the market also thinks 51% control is destructive, then it is individually rational to come on board and also sell. The speculator profits all the way down until the new equilibria is reached and bitcoin lies in ruin. The mystery miner and the speculator could be one and the same person. Due to this self-fulfilling prophecy effect, the actual threshold which leads to 51% control is much less than 51%, it could potentially be 20 or 30% though it is very hard to guess because it depends on perceptions about the price consequences of 51% attack.
This is not FUD, just textbook international finance. If you think those are one and the same, I’m selling tinfoil hats on bitmit.
Right now with block reward at 50, an attack is as costly as it will ever be. Around December, monopoly gets easier, and the plan for the future is to make it even easier.
Shouldn’t the development team make some statement about how they plan to address this pressing issue?
It’s percentages, so it’s not “only 3.4 times more,” but more like 6 to 7 times more. And if you want guaranteed 51%+ control, more like 10 times more than the 15%
Let’s start out ignoring any possible speculative attack.
If we apply your logic, solo miner needs to increase by less than 6-fold to go from 15% to 51%. So it might appear that I overstated things (I did to some small degree). However, your logic assumes that GPU miners never respond to difficulty increases by shutting down there rigs. This is obviously untrue, so the actual increase necessary will be less than 6-fold. If we assume that at least 41% of incumbent miners are near their shutdown threshold (that is will shutdown if difficulty increases), then the actual number is 3.4 as originally stated. 3.4 is a lower bound, around 5.9 is an upper bound. I think 3.4 is closer to the real answer.
Let’s introduce speculative attackIn addition, if we allow for rational forward-looking behavior on the part of miners, coupled with perceived catastrophe in the event of 51% attack, then the actual number could be much less than 3.4, perhaps even a 2-fold increase could be enough to cause the entire system to unravel.
Care to plug in costs into those estimates? Currently they are very likely earning a profit. If they expand by 3.4 times, not only will they have to spend on hardware, all of their own hardware will become quite unprofitable, too. After they get to 51%, they will still have to deal with a long unprofitable lag as other miners continue to exit and profitability takes time to decrease. Unless the purpose is to destroy Bitcoin (a very expensive task), this seems like a pretty stupid thing to do.
Btw, Gavin and group have figured out that this 15% is coming from a botnet, so I guess none of this is even an issue.
don’t bother, guy doesn’t know what he’s talking about. Even in my math above that assumes the network gains ZERO except for one miner suddenly. In fact the order of power gain would be higher because it would have to make up for the rest of the network gain.
(attackerOriginalRate+attackerRateGain+otherNetworkGain) / (networkOriginalRate+attackerRateGain+otherNetworkGain) = .51
A+B+C / D+B+C = .51
A+B+C = .51D+.51B+.51C
B = .51D-A+.51B-.49C
-.49B=.51D-A-.49C
B = -.49(.51D-A-.49C)
B = .49A – .2499D + .2401C
plug in any values for original rates and what the network gains to see how insane it would be to suddenly be 51% of the network. Nope.
don’t pretend like you had some other logic besides 15*3.4 = ~51… which is just terrible math.
if you are 15 out of 100 computing power, that is 15%.
to become 51% of the network would be
(yourRate+newRate) / (totalRate+newRate)=.5
so
15+x
——–
100+x = .51
15+x = 51+.51x
.49x = 36
x = 73.46938
you would have to go from 15 power to 88.46938 power, with the network total going from 100 to 173.46938 power.
86.4285/171.4285 = .51
86.4285 / 15 = 5.76x
so yeah, 51% is not 3x as hard, its 5.76x as hard. You can backpedal all you want.
Less than 6-fold when everyone else in the world has access to similar technology? Laff. Not even google was 1-2x more powerful than the previous search engine.
in that case it is each miner’s duty to try to occupy as much of the network mining power as possible, if you’re all mining just as hard as the mystery miner it would be extremely difficult to overtake the rest.