5 ms·

Proof of work has always had an economic flaw that you could theoretically temporarily rent enough mining power to perform double spends of more value than the

Proof of work has always had an economic flaw that you could theoretically temporarily rent enough mining power to perform double spends of more value than the cost of renting those devices. But this attack has never been performed because the reality of all these cryptocurrencies is that the security depends only relatively weakly on proof of work. Instead it relies on trust between the main stakeholders: miners, big nodes and developers. This is just like any other human organisation. That trust is only reinforced by proof of work, making it easier for new parties to become trusted.

crazypython 5y ago
https://www.crypto51.app/ https://www.crypto51.app/ To execute a double spend, you the one sending the transaction and the miner must coordinate. For large transactions, it is recommended to wait for six confirmations. (six blocks that agree with the transaction and have not been 51%ed.) The 1 hour 51% cost of Bitcoin is 1.9m$. However, you would need much more time than that to find six consecutive blocks alone, without the help of the network. So, while the network is 6 blocks ahead, you need to find 7 blocks. The network moves forward a block, you must move forward more than one block to catch up. This could take a long time, and longer the more confirmations required- each confirmation makes each previous transaction exponentially more secure. Simply controlling the mining power momentarily only puts recent transactions vulnerable. However, that much hardware is available for rent-see “Nicehashable.”
- yuliyp 5y ago
  I don't see how more confirmations makes it exponenrially harder to mount a 51% attack: you just need to be mining faster privately than the remaining community until you (a) have a lead, and (b) have maintained that lead for the confirmation window.
  - yanmaani 5y ago
    There's luck involved too. In the limiting case, imagine you have 0.001% of the network hash rate and need a 1 block lead. This can happen every now and then, but getting a 2 block lead is basically impossible. The exact formula is given in the Bitcoin whitepaper <https://bitcoin.org/bitcoin.pdf https://bitcoin.org/bitcoin.pdf>, see page 6 and on.
    - yuliyp 5y ago
      According to that formula, in a 51% attack, probability of success = 1 (since 51% attack means you are more likely to find a hash than all others combined), right?
  - bellajbadr 5y ago
    That's called Selfish mminign but the prob is to be faster than the network!
xiphias2 5y ago
Renting out equipment for miners is dangerous, as an attack on the network can wipe out all their sunk costs. Bitcoin miner farms are very strongly incentivized to be aligned with the long term health of the system, so they try to play nice close to the ethos of Bitcoin (they learned enough from the blocksize wars that they can get burned if they go against Bitcoin)
- Eelongate 5y ago
  An attack on bitcoin might have other motivations. For instance, a powerful government might attack bitcoin specifically to undermine bitcoin's credibility and utility.
  - xiphias2 5y ago
    I was talking about motivations for renting out, not attacking. A goverment would just take over the miners with violence probably if it wanted to attack the system.
- paulgb 5y ago
  If a halving forces a bunch of miners off the market (because their expected revenue drops below operating cost), there could be a huge glut of hardware flooding the market that is not profitable for honest mining but allows a bad actor to temporarily accumulate hashrate.
  - xiphias2 5y ago
    Bitcoin's price goes up faster than the effect of the halving (16x vs 0.5x every 4 years), and with the current chip shortage and slowdown of improvements miners generally plan for 4 years, but old, inefficient miners with low utilization can become an attack vector.
    - adgjlsfhk1 5y ago
      If your security model requires exponential increase in the value of crypto, it's a bad security model.
xg15 5y ago
> Instead it relies on trust between the main stakeholders: miners, big nodes and developers. In other words, it relies not on one central authority but on a small clique of central authorities.
betwixthewires 5y ago
This attack has happened, at least twice to ETC and many others. Proof of work networks with the same hash algorithm are a threat to one another, particularly, if a network exists that is profitable enough to have exorbitant resources dedicated to mining, those resources are available to attack a much smaller network if that becomes more profitable for some period than just continuing to mine the bigger one. Proof of work then only protects the largest projects using unique hash algorithms.
RF_Savage 5y ago
My understanding was that this was demoed with some small coin by just hiring the 51% on niceHash for a few hundred usd. Edit: I was lookimg for the artivle I was referring to and saw that Ethereoum Classic (ETC) has been attacked simce via just that method. https://www.coindesk.com/markets/2020/08/29/ethereum-classic-hit-by-third-51-attack-in-a-month/ https://www.coindesk.com/markets/2020/08/29/ethereum-classic...
celeduc 5y ago
51% attacks have been performed on some smaller PoW chains, specifically Ethereum Classic. It's not that hard. https://www.crypto51.app/ https://www.crypto51.app/
runeks 5y ago
> Proof of work has always had an economic flaw that you could theoretically temporarily rent enough mining power to perform double spends of more value than the cost of renting those devices. It’s not a flaw if it’s only theoretical. In practice, no miner with billions of dollars of capital bound in mining hardware would rent it out to someone who might do something that would significantly depreciate this capital (e.g. attack the Bitcoin network).

crazypython 5y ago

https://www.crypto51.app/ https://www.crypto51.app/ To execute a double spend, you the one sending the transaction and the miner must coordinate. For large transactions, it is recommended to wait for six confirmations. (six blocks that agree with the transaction and have not been 51%ed.) The 1 hour 51% cost of Bitcoin is 1.9m$. However, you would need much more time than that to find six consecutive blocks alone, without the help of the network. So, while the network is 6 blocks ahead, you need to find 7 blocks. The network moves forward a block, you must move forward more than one block to catch up. This could take a long time, and longer the more confirmations required- each confirmation makes each previous transaction exponentially more secure. Simply controlling the mining power momentarily only puts recent transactions vulnerable. However, that much hardware is available for rent-see “Nicehashable.”

yuliyp 5y ago
I don't see how more confirmations makes it exponenrially harder to mount a 51% attack: you just need to be mining faster privately than the remaining community until you (a) have a lead, and (b) have maintained that lead for the confirmation window.
- yanmaani 5y ago
  There's luck involved too. In the limiting case, imagine you have 0.001% of the network hash rate and need a 1 block lead. This can happen every now and then, but getting a 2 block lead is basically impossible. The exact formula is given in the Bitcoin whitepaper <https://bitcoin.org/bitcoin.pdf https://bitcoin.org/bitcoin.pdf>, see page 6 and on.
  - yuliyp 5y ago
    According to that formula, in a 51% attack, probability of success = 1 (since 51% attack means you are more likely to find a hash than all others combined), right?
- bellajbadr 5y ago
  That's called Selfish mminign but the prob is to be faster than the network!

xiphias2 5y ago

Renting out equipment for miners is dangerous, as an attack on the network can wipe out all their sunk costs. Bitcoin miner farms are very strongly incentivized to be aligned with the long term health of the system, so they try to play nice close to the ethos of Bitcoin (they learned enough from the blocksize wars that they can get burned if they go against Bitcoin)

Eelongate 5y ago
An attack on bitcoin might have other motivations. For instance, a powerful government might attack bitcoin specifically to undermine bitcoin's credibility and utility.
- xiphias2 5y ago
  I was talking about motivations for renting out, not attacking. A goverment would just take over the miners with violence probably if it wanted to attack the system.
paulgb 5y ago
If a halving forces a bunch of miners off the market (because their expected revenue drops below operating cost), there could be a huge glut of hardware flooding the market that is not profitable for honest mining but allows a bad actor to temporarily accumulate hashrate.
- xiphias2 5y ago
  Bitcoin's price goes up faster than the effect of the halving (16x vs 0.5x every 4 years), and with the current chip shortage and slowdown of improvements miners generally plan for 4 years, but old, inefficient miners with low utilization can become an attack vector.
  - adgjlsfhk1 5y ago
    If your security model requires exponential increase in the value of crypto, it's a bad security model.

xg15 5y ago

> Instead it relies on trust between the main stakeholders: miners, big nodes and developers. In other words, it relies not on one central authority but on a small clique of central authorities.

betwixthewires 5y ago

This attack has happened, at least twice to ETC and many others. Proof of work networks with the same hash algorithm are a threat to one another, particularly, if a network exists that is profitable enough to have exorbitant resources dedicated to mining, those resources are available to attack a much smaller network if that becomes more profitable for some period than just continuing to mine the bigger one. Proof of work then only protects the largest projects using unique hash algorithms.

RF_Savage 5y ago

My understanding was that this was demoed with some small coin by just hiring the 51% on niceHash for a few hundred usd. Edit: I was lookimg for the artivle I was referring to and saw that Ethereoum Classic (ETC) has been attacked simce via just that method. https://www.coindesk.com/markets/2020/08/29/ethereum-classic-hit-by-third-51-attack-in-a-month/ https://www.coindesk.com/markets/2020/08/29/ethereum-classic...

celeduc 5y ago

51% attacks have been performed on some smaller PoW chains, specifically Ethereum Classic. It's not that hard. https://www.crypto51.app/ https://www.crypto51.app/

runeks 5y ago

> Proof of work has always had an economic flaw that you could theoretically temporarily rent enough mining power to perform double spends of more value than the cost of renting those devices. It’s not a flaw if it’s only theoretical. In practice, no miner with billions of dollars of capital bound in mining hardware would rent it out to someone who might do something that would significantly depreciate this capital (e.g. attack the Bitcoin network).