How much would it cost for a government to undermine Bitcoins?

11

1

Articles by for example the Economist stating that Bitcoins are being used for drug trading (Silk Road), I believe that the biggest risk to Bitcoins is that a government will pull a stunt similar to Stuxnet. This was also the answer of David Schwartz to a previous question of mine.

Therefore, I believe that we should make a metric, similar to Bus Factor for software projects, which indicates risk of such an event with a single number. I propose that the metric should be called CSU (Cost of Secret Undermining). Later we could calculate the cost of undermining in other ways, like by banning businesses that accept bitcoins and so on, but that will be much harder.

Therefore, my question is: What is the current lowest CSU (Cost of Secret Undermining). This means, what is the cheapest way for someone (for example a government) to secretly undermine Bitcoins to the extent that Bitcoins are no longer useful as a medium of exchange, for example by utilizing the method suggested by David Schwartz. We should compare the number that we find with the current budget of the US fight against illegal drugs. Of course, the expected impact on the fight against illegal drugs is a deciding factor. Even though it will be very hard, we should try to estimate the development time needed to launch such an attack as this will be a cost the attacker would have to calculate with.

This question is related to this question, but now I would like to know the cost of undermining Bitcoins instead of just launching one single 51% attack.

David

Posted 2012-10-07T11:14:41.410

Reputation: 3 374

Would this be a black project or one that is publicly acknowledged as being carried out? It makes a big difference in the cost. Also, would it be just one country or would it be a cartel of nations carrying out such an attack? – Stephen Gornick – 2012-10-08T02:34:44.723

@StephenGornick A black one. The costs should then not be significantly affected by whether the undermining is carried out by a single or several organizations or governments. – David – 2012-10-08T05:36:12.070

It sure can. If this hypothetical black ops deal requires cooperation from a foreign government (e.g., block port 8333 traffic) then there is more paid in bribe money than would be the cost of making an official request. – Stephen Gornick – 2012-10-10T20:12:55.420

Answers

10

As I suggested in my Bitcoin master thesis, a 51% attack would cost in a bulk part of about 7 million dollars to execute. The figure will change depending on the technology used, the current difficulty and so forth. This option would make Bitcoin or any other Bitcoin-like currency useless.

However, if the USA goverment was to crack down on Bitcoin as a terrorist activity, it could detain any Bitcoin developer in the country indefinitely, which would probably be cheaper. This option would most likely generate a lot of backlash from the community, and could encourage other programmers to take over the project.

ThePiachu

Posted 2012-10-07T11:14:41.410

Reputation: 42 235

2As you describe on page 42 in your thesis, those 7 million dollars only include the price of graphics cards, not other hardware needed, nor development costs of exploiting the 51% attack. Nor does it include variations in hashrate in the network, which might make it much cheaper to launch such an attack at certain times. Also, it might be possible to rent such hardware instead of buying it, which would make it significantly cheaper. If we said that the final price of launching such an attack would be 15-20 million, it really is peanuts for a government fighting illegal drugs. – David – 2012-10-07T15:21:03.370

@David, the cost is not much but it's still big enough and the task is too complicated to execute this attack in secrete, therefore US government would need a solid political agenda for that. – Serith – 2012-10-07T17:03:39.877

@Serith I would argue that this would be peanuts compared to developing Stuxnet. And it was certainly developed in secrecy. That was truly an amazing effort and I would guess that it involved much larger sums than what we are talking about here. – David – 2012-10-07T19:12:56.427

@David Well, I did say that it is in the bulk part of 7 million - as in probably within one order of magnitude. Probably you could get it cheaper actually if you used that as R&D money to create some high-end ASIC devices and then produce them in bulk. But yeah, I'm not a specialist at estimating such costs as I have no experience in it. – ThePiachu – 2012-10-07T19:29:13.170

3@Serith It wouldn't be too complicated to execute it in secrecy - 7 million is not that big of an expenditure in comparison to some military budget. You could introduce the machines gradually as some solo miners in various locations, and when you reach critical mass turn them from honest miners to attackers and I don't think anyone would expect it. – ThePiachu – 2012-10-07T19:33:10.497

2

The NSA, as well as others, already have the computing power to achieve this. I'm sure no one but the NSA know's where the processing power of this fully armed and operational super computer is aimed.

– Matt – 2012-10-07T22:22:09.077

@mindthemonkey Actually, those computers might not be the best choice to use against Bitcoin. Generally, supercomputers are used for fancy operations, like using floating point numbers and so forth, while Bitcoin just uses raw integers, making GPUs blow CPUs out of the water. – ThePiachu – 2012-10-08T05:11:07.640

@ThePicahu Correct for most, but that doesn't mean they can't achieve it. If you check down the list you will see a number of them are based on Telsa 20 series coproccesors including the US. Even Google, Amazon, Intel, Ibm, Sun/Oracle, Microsoft probably have the compute power to achieve this.

– Matt – 2012-10-08T07:58:48.960

1I would guess that it would cost a lot more than $7 million for the US government to crack down on Bitcoin as a terrorist activity. – JimN – 2012-10-13T04:50:44.990

3

In certain situations it gets very easy: If the NSA decides to send national security letters to developers they can be forced to do anything without being allowed to talk about it. Forcing them to 1: publish low quality updates and introduce backdoors, 2: dump-sell their Bitcoin over a predefined period should be quite effective.

This could make Bitcoins effectively useless in the short term. It might not work if the Americans don't manage to get a large enough part of the network to dump-sell, but I think there's sufficient wealth to push it below the cent range. Subsequently the crappy development should help it die a slow death.

This has a negligible cost.

Frankly I think becoming a horrible part of the community would be the most effective.

Team: 3 programmers 2 generic agents

The programmers create slow and abusive parts of the community software and have them be "exploited" or "hacked" frequently. Pushing media not to publish about Bitcoin, but when they do publish about these failures. They also continually publish expensive-to-process and large transactions in order to make running Bitcoin more expensive.

The generic agents attempt to shut down any Bitcoin businessman and/or buy out Bitcoin businesses and corrupt them.

There's also fake adverts created on Bitcoin trading sites, trying hard as they can to exploit the reputation systems.

These measures could make Bitcoin an extremely horrible-to-use system and should give it little extra exposure.

Naturally cracking existing Bitcoin businesses would also help.

Simply degrading the trust in Bitcoin whilst minimizing exposure would undermine Bitcoin. I do not think that secret undermining can stop ALL Bitcoin transactions.

This has a cost equal to 5 agents continuously.

Lodewijk

Posted 2012-10-07T11:14:41.410

Reputation: 1 527

2This is not secret undermining. – David – 2012-10-08T05:41:38.667

It's stealthy. Can't destroy something without leaving traces so I figure it's about right. David's methods are also quite detectable, if you'd try. Do give an example of what "secret undermining" really should be. – Lodewijk – 2012-11-27T14:40:36.047