3 ms·

I have a hard time believing that scaling back the hash rate gets you vulnerable. $219 million dollars per year ($7/second) in cryptographic operations to supp

I have a hard time believing that scaling back the hash rate gets you vulnerable. $219 million dollars per year ($7/second) in cryptographic operations to support a maximum of 7 transactions per second - the bitcoin network's current maximum is... a bit ridiculous

danbruc 12y ago
The hardware is out there. If you reduce the hash rate and therefore the operation costs tenfold you end up with a 25 PHash/s network on the one hand and unused hardware worth 225 PHash/s on the other hand. Significantly reducing the network hash rate without getting vulnerable seems a pretty hard problem to me. I did not make the numbers up - hash rate and transaction volume [1], mining efficiencies [2] and energy costs [3]. Actually I think I underestimated the costs quite a bit. [1] https://blockchain.info/stats https://blockchain.info/stats [2] https://en.bitcoin.it/wiki/Mining_hardware_comparison https://en.bitcoin.it/wiki/Mining_hardware_comparison [3] http://www.statista.com/statistics/263492/electricity-prices-in-selected-countries/ http://www.statista.com/statistics/263492/electricity-prices...
- logicallee 12y ago
  Are you saying it's worth the time of the owners of the hardware to break bitcoin? Why? When a mining pool recently reached 50% of hashrate, they didn't use it to break bitcoin. Just because something isn't cryptographically secure anymore (e.g. 51% owned by 1 single party) doesn't break anything. Likewise I doubt the $200 million worth of hardware needs to be used on the hashrate.
  - danbruc 12y ago
    The cheapest hardware listed on the page linked before is 2.5 GHash/s/$. This yields hardware worth 100 million Dollars as a lower bound. But I did not think so much about hardware costs when I said that it is probably underestimated but because I picked the efficiency from the top of the range and energy costs from the bottom. And it does not necessarily break anything but the security of Bitcoin is build on the assumption that it is very hard for any single entity to posses the majority of the hash rate. If you take the 100 million Dollars of hardware, reducing the hash rate tenfold will make the network attackable with only 10 million Dollars which seems pretty cheap to me. And being vulnerable does not imply the attack is actually executed, it only means it could happen anytime. But do you want your money in a system that could be attack anytime?
    - logicallee 12y ago
      As I mentioned, 51% of the network did get controlled by one single entity. The question "do you want your money in a system that could be attacked anytime" has two answers: "Yes" or "I don't want any money." There is no perfectly secure third alternative.
      - danbruc 12y ago
        They controlled 51 % and decided not to abuse their power. But the idea behind Bitcoin is that you don't have to trust any single entity and if you allow this to happen this idea goes straight out of the window. You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet and hope they will not abuse your trust.
        kanzure 12y ago
        > You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet Hmm, that's not really how these attacks work. When a miner doesn't mine according to the rules, their blocks are invalid and rejected by the network. edit: Unfortunately you have now given me an opportunity to admit that Excel is probably capable of running Bitcoin-equivalent software, though.... I would prefer not thinking about this.
        ikeboy 12y ago
        There was a 51% attack against a gambling site before http://arstechnica.com/security/2014/06/bitcoin-security-guarantee-shattered-by-anonymous-miner-with-51-network-power/ http://arstechnica.com/security/2014/06/bitcoin-security-gua...
        logicallee 12y ago
        But you don't actually mean "You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet and hope they will not abuse your trust." You think you mean this, but you simply don't. It's a nice ideal but none of it is true. The broken system we have, where 51% decided not to abuse their power, works. A trivial replacement without any cryptography relying on the honor system demonstrably (and obviously) doesn't. That is not an insignificant difference. While you would like the cryptography to be provably secure, it simply isn't (51% control-by-one-group actually happened), and it wasn't a disaster. It was still far better than the alternative. The world is not black and white.

danbruc 12y ago

The hardware is out there. If you reduce the hash rate and therefore the operation costs tenfold you end up with a 25 PHash/s network on the one hand and unused hardware worth 225 PHash/s on the other hand. Significantly reducing the network hash rate without getting vulnerable seems a pretty hard problem to me. I did not make the numbers up - hash rate and transaction volume [1], mining efficiencies [2] and energy costs [3]. Actually I think I underestimated the costs quite a bit. [1] https://blockchain.info/stats https://blockchain.info/stats [2] https://en.bitcoin.it/wiki/Mining_hardware_comparison https://en.bitcoin.it/wiki/Mining_hardware_comparison [3] http://www.statista.com/statistics/263492/electricity-prices-in-selected-countries/ http://www.statista.com/statistics/263492/electricity-prices...

logicallee 12y ago
Are you saying it's worth the time of the owners of the hardware to break bitcoin? Why? When a mining pool recently reached 50% of hashrate, they didn't use it to break bitcoin. Just because something isn't cryptographically secure anymore (e.g. 51% owned by 1 single party) doesn't break anything. Likewise I doubt the $200 million worth of hardware needs to be used on the hashrate.
- danbruc 12y ago
  The cheapest hardware listed on the page linked before is 2.5 GHash/s/$. This yields hardware worth 100 million Dollars as a lower bound. But I did not think so much about hardware costs when I said that it is probably underestimated but because I picked the efficiency from the top of the range and energy costs from the bottom. And it does not necessarily break anything but the security of Bitcoin is build on the assumption that it is very hard for any single entity to posses the majority of the hash rate. If you take the 100 million Dollars of hardware, reducing the hash rate tenfold will make the network attackable with only 10 million Dollars which seems pretty cheap to me. And being vulnerable does not imply the attack is actually executed, it only means it could happen anytime. But do you want your money in a system that could be attack anytime?
  - logicallee 12y ago
    As I mentioned, 51% of the network did get controlled by one single entity. The question "do you want your money in a system that could be attacked anytime" has two answers: "Yes" or "I don't want any money." There is no perfectly secure third alternative.
    - danbruc 12y ago
      They controlled 51 % and decided not to abuse their power. But the idea behind Bitcoin is that you don't have to trust any single entity and if you allow this to happen this idea goes straight out of the window. You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet and hope they will not abuse your trust.
      - kanzure 12y ago
        > You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet Hmm, that's not really how these attacks work. When a miner doesn't mine according to the rules, their blocks are invalid and rejected by the network. edit: Unfortunately you have now given me an opportunity to admit that Excel is probably capable of running Bitcoin-equivalent software, though.... I would prefer not thinking about this.
      - ikeboy 12y ago
        There was a 51% attack against a gambling site before http://arstechnica.com/security/2014/06/bitcoin-security-guarantee-shattered-by-anonymous-miner-with-51-network-power/ http://arstechnica.com/security/2014/06/bitcoin-security-gua...
      - logicallee 12y ago
        But you don't actually mean "You could just as well throw out all the expensive cryptographic stuff, send them your transactions and let them manage them in an Excel sheet and hope they will not abuse your trust." You think you mean this, but you simply don't. It's a nice ideal but none of it is true. The broken system we have, where 51% decided not to abuse their power, works. A trivial replacement without any cryptography relying on the honor system demonstrably (and obviously) doesn't. That is not an insignificant difference. While you would like the cryptography to be provably secure, it simply isn't (51% control-by-one-group actually happened), and it wasn't a disaster. It was still far better than the alternative. The world is not black and white.