8 ms·

> Significantly, the Xinjiang and Sichuan regions of China combined account for more than 50 percent of the overall Bitcoin mining hash rate I don't get it, do

> Significantly, the Xinjiang and Sichuan regions of China combined account for more than 50 percent of the overall Bitcoin mining hash rate I don't get it, doesn't this effectively mean China can force a 51% attack by just strongarming the pools that operate there?

CharlesW 5y ago
Yes.
- mrits 5y ago
  I'm not sure about that. Most miner operations in the US run only a fraction of the day to take advantage of electrical savings that a lot of operations in China wouldn't be doing. So if miners started returning higher yields we would output a larger hash rate. If we actually needed to protect against a 51% attack we could ramp up accordingly.
  - gruez 5y ago
    The difficulty adjustment only happens every 2016 blocks (~2 weeks), so the higher yields wouldn't happen unless they're down for an extended period of time.
    - mrits 5y ago
      That is correct but I don't see why you are pointing it out.
      - gruez 5y ago
        It's a reply at this part of GP's comment: >So if miners started returning higher yields we would output a larger hash rate
        mrits 5y ago
        That statement was to show why there is room to increase hash rate and not how you'd defend against a 51% attack. Most of the miners in the US are not mining for a significant portion of the day.
    - ndr 5y ago
      How much longer than 2 weeks will it take given the sudden drop in miners?
      - gruez 5y ago
        It depends on how long ago the previous retarget is and the current hashrate. The retarget is based on blocks, so if the retarget was 1008 blocks ago, but hashrate dropped by 50%, then it'll take 2 weeks until the next retarget. There's a chart here with the current estimates: https://fork.lol/pow/retarget https://fork.lol/pow/retarget
ChainOfFools 5y ago
yeah, well more specifically they could force a preferred fork to assume the "Bitcoin" ticker (BTC) by denying power to chinese miners on the disfavored fork. exchanges and miners, rather than the developers, ultimately determine this, as there's no in-band way to control what ticker is used for trading what token.
- gruez 5y ago
  That was tried a few years ago with the segwit2x fork. A group of miners comprising 82% of the hashpower got together and agreed to activate the segwit2x fork. Despite having a overwhelming majority of the hashpower the fork was never happened due to lack of community consensus. https://dcgco.medium.com/bitcoin-scaling-agreement-at-consensus-2017-133521fe9a77 https://dcgco.medium.com/bitcoin-scaling-agreement-at-consen...
  - ChainOfFools 5y ago
    the consensus vote to activate sw2x on existing BTC network was an in-band mechanism, and set to some super impossible threshold like 95 or 99% to activate a protocol change. this is a different situation, where an out of band entity can selectively shut off a significant fraction of hashpower, and unlike the weedy obscurity of an architectural spat between largely unknown (by the mainstream) such a move would (indeed, will) make major waves in mainstream Joe Pension Fund/ Number Go Up news, who will not be familiar with nor interested in the technical arguments for the fork that is being actively suppressed. there's no weeks-long debate window here. the stakes are very different now than in the Jihan boogeyman days, you can't handwave away this critical risk to stability or spin it as technical trivia of no concern to non-specialists.
    - gruez 5y ago
      >the consensus vote to activate sw2x on existing BTC network was an in-band mechanism, and set to some super impossible threshold like 95 or 99% to activate a fork. Source for this? I searched around and the target was 80%. https://www.weusecoins.com/what-is-segwit2x/ https://www.weusecoins.com/what-is-segwit2x/ > such a move would (indeed, will) make major waves in mainstream Joe Pension Fund/ Number Go Up news, who will not be familiar with nor interested in the technical arguments for the fork that is being actively suppressed. there's no weeks-long debate window here. Fair point, given how the markets reacted to this drop in hashrate.
rantwasp 5y ago
that’s not how 51% attacks work. you have the miners but you also have the actual bitcoin nodes that are witnessing the blockchain growth.
- bonestamp2 5y ago
  Do the witnesses have any weight on the consensus?
  - ric2b 5y ago
    Depends on who is using them. If it's a large business/exchange: definitely.
dnautics 5y ago
No, it means that theoretically a 51% is possible but the levels of coordination required to do that and the ability of the bitcoin network to figure it out and ignore it are means that it is unlikely to be effective.
- sillysaurusx 5y ago
  Plus, the model of BTC as “digital gold” is pretty resilient to 51% attacks: if you don’t touch your coins, no one can take them away from you. (A 51% attack can’t reassign coins.) Of course, the price of BTC might tank in response to the news of a successful 51% attack, but it’ll probably bounce back just like every other event in BTC history.
  - paulmd 5y ago
    > if you don’t touch your coins, no one can take them away from you. (A 51% attack can’t reassign coins.) it can't generate new unsigned transactions but it can roll back the network to a state when someone else owned those coins. from a practical perspective there is no difference, you no longer have those coins.
    - fouc 5y ago
      there's a limit to how far it can roll back the network, it probably wouldn't be able to roll back more than a few days..
      - yuvadam 5y ago
        "Rolling back" the network means successfully creating a new longest chain that forks off of an arbitrary old block. The further back in time you to "rollback" to, the less probable it is for the 51% to successfully mine all those needed blocks to get up to par with the existing longest chain. Realistically speaking, this sort of attack is highly improbable.
        sillysaurusx 5y ago
c54 5y ago
> doesn't this effectively mean China can force a 51% attack by just strongarming the pools that operate there? Yes, in theory. There's a coordination problem to get all the pools to do what you want, but if they (or anyone) could pull that off, then yes. So then what happens? If a single entity controls >50% of the hash power of the network, they could start finding blocks with arbitrary (invalid) transactions and adding them to the chain. The transactions they'd be able to insert are called 'double-spends', which is that wallet A sends its balance to wallet B, and then in the next block sends the same balance (magicked into existence) to some third wallet C. All this can happen. The way the network is supposed to work though, miners don't just commit new blocks and call it a day, the 'full nodes' have to accept the blocks. The miners are more or less offering up solved blocks to the full nodes for their acceptance. As a full node it's trivial to take a look at this double-spend from A->C and say 'nope this doesn't look like a valid block to me, try again'. At this point if enough full nodes do this, the 51% of 'bad' miners are just wasting their time, the true hashrate of the network is now 49% of what it was a moment ago, and the remaining 'good' miners will continue to find blocks. Business as usual. It's a much harder problem to control >51% full nodes, mostly because they're more geographically distributed. This is because they're pretty simple to set up at home and run, without worrying about up front cost of miners, electricity and noise and so on. A bitcoin full node can run on a raspi with a 1tb external hard drive, and a normal consumer internet connection. Another interesting tidbit is that with control of 51% of the miners, this still only means a malicious actor finds a block first 51% of the time... to get the 6 blocks usually required for confirmation (in a row) is a ~3% chance. Not a perfect safeguard, but just some perspective on how hard it is to execute an attack like this. The idea is that the miners are beholden to the full nodes, who are the actual actors in the distributed consensus system. There's some other game theory around why one would want to do obvious double spends after having that much control (why not just sell the bitcoin and take profits?). But there are definitely reasons a state actor would want to try to destroy the network.
- dangero 5y ago
  inserting invalid transactions is not the double spend issue. If their hash rate is over 50% they can "unconfirm" a confirmed transaction by making a chain with more work on it. This is how the attack plays out in simplified form: Bitcoin is sent to an exchange by the attacker, confirmed, sold and withdrawn for cash. Attacker then makes a chain with more work on it than the chain where the Bitcoin was sent to the exchange. Now as far as Bitcoin network is concerned, the coins were never sent to the exchange, but the exchange already let the user sell the Bitcoin and cash out. So now the exchange has neither the cash or the Bitcoin. The attacker has both.
jabbany 5y ago
In theory yes, but in practice it's rather infeasible to do 51% attacks IRL --- the main constraint is the people not the math. A 51% attack mainly lets you double spend, so you have to: (1) get a lot of coins and spend them on something (2) start mining a fork where that spending didn't happen (but other real transactions continue to happen) --- your hash power will now need to be dedicated to this new fork at the exclusion of the old "real" fork since all future hashes will diverge (3) wait until the thing you paid for with coins can no longer be "taken back", all while continuing efforts on the alternate fork (4) do the 51% attack to swap out the old "real" fork with your new fork where you never bought the thing, thus keeping the thing _and_ the coins you spent on them (5) all this needs to be balanced with the opportunity cost had you just continued mining on the real chain... Such an attack would trigger immediate attention since at (2) the main chain would suddenly see a drop in contributions. Also realistically most things you can purchase with coins (like fiat or other goods/services) can be relatively easily "taken back" unless a considerable amount of time is spent to launder them --- say if you "buy" a car with crypto and reverse the transaction, what you've really done is just stolen a car with a lot of extra steps... Which means your only real viable path is to redeem for some other cryptocurrency... Now given that a big attack just happened on the largest market cap crypto, there's a pretty good chance the other coins are going to take a hit and exchanges might even temporarily limit activities while this situation is "investigated", greatly reducing what you can actually do at the end of the day. This kind of thing is more reasonable on the smaller market cap "meme" coins where you can just exchange them for for a mass-market crypto.
- roofwellhams 5y ago
  What if this drop in hash is to use on the 2nd chain?
  - Thorrez 5y ago
    Well we have a good explanation of the drop: the power outages. Also with 25% hash power you're going to have a hard time doing a 51% attack.
    - 5y ago
emerged 5y ago
I like how the replies are either yes or “yes, but <12 paragraphs of but>”
Zamicol 5y ago
Yes.
sashimi-houdini 5y ago
Yes it does. And if/when a large enough player finds it in their interest to mount such an attack, they might very well do it. See https://www.crypto51.app/ https://www.crypto51.app/ for a (somewhat outdated) price chart. Remember the first law of cryptocurrency though: "in case of emergency, betray your core principles." In the event of such an attack, the rest of the system (miners, developers) may be able to simply declare the "evil" side of the fork "untrue" and hardcode the other fork as true. The fact that this is exactly opposite to Bitcoin's stated goals is no problem at all. See e.g. Ethereum's response to the "DOA Attack" as an example.
usrusr 5y ago
I've been wondering about possible subtle benefits hashrate dominance could bring other than double spending. I'd be surprised if there wasn't some way to benefit nicely from being able to load the dice of which transactions are committed and which have to wait for some minority miner to win the race? Certainly nothing on the scale that it would pay for the hashrate by itself, but that's already covered by regular fees and reward.

CharlesW 5y ago

Yes.

mrits 5y ago
I'm not sure about that. Most miner operations in the US run only a fraction of the day to take advantage of electrical savings that a lot of operations in China wouldn't be doing. So if miners started returning higher yields we would output a larger hash rate. If we actually needed to protect against a 51% attack we could ramp up accordingly.
- gruez 5y ago
  The difficulty adjustment only happens every 2016 blocks (~2 weeks), so the higher yields wouldn't happen unless they're down for an extended period of time.
  - mrits 5y ago
    That is correct but I don't see why you are pointing it out.
    - gruez 5y ago
      It's a reply at this part of GP's comment: >So if miners started returning higher yields we would output a larger hash rate
      - mrits 5y ago
        That statement was to show why there is room to increase hash rate and not how you'd defend against a 51% attack. Most of the miners in the US are not mining for a significant portion of the day.
  - ndr 5y ago
    How much longer than 2 weeks will it take given the sudden drop in miners?
    - gruez 5y ago
      It depends on how long ago the previous retarget is and the current hashrate. The retarget is based on blocks, so if the retarget was 1008 blocks ago, but hashrate dropped by 50%, then it'll take 2 weeks until the next retarget. There's a chart here with the current estimates: https://fork.lol/pow/retarget https://fork.lol/pow/retarget

ChainOfFools 5y ago

yeah, well more specifically they could force a preferred fork to assume the "Bitcoin" ticker (BTC) by denying power to chinese miners on the disfavored fork. exchanges and miners, rather than the developers, ultimately determine this, as there's no in-band way to control what ticker is used for trading what token.

gruez 5y ago
That was tried a few years ago with the segwit2x fork. A group of miners comprising 82% of the hashpower got together and agreed to activate the segwit2x fork. Despite having a overwhelming majority of the hashpower the fork was never happened due to lack of community consensus. https://dcgco.medium.com/bitcoin-scaling-agreement-at-consensus-2017-133521fe9a77 https://dcgco.medium.com/bitcoin-scaling-agreement-at-consen...
- ChainOfFools 5y ago
  the consensus vote to activate sw2x on existing BTC network was an in-band mechanism, and set to some super impossible threshold like 95 or 99% to activate a protocol change. this is a different situation, where an out of band entity can selectively shut off a significant fraction of hashpower, and unlike the weedy obscurity of an architectural spat between largely unknown (by the mainstream) such a move would (indeed, will) make major waves in mainstream Joe Pension Fund/ Number Go Up news, who will not be familiar with nor interested in the technical arguments for the fork that is being actively suppressed. there's no weeks-long debate window here. the stakes are very different now than in the Jihan boogeyman days, you can't handwave away this critical risk to stability or spin it as technical trivia of no concern to non-specialists.
  - gruez 5y ago
    >the consensus vote to activate sw2x on existing BTC network was an in-band mechanism, and set to some super impossible threshold like 95 or 99% to activate a fork. Source for this? I searched around and the target was 80%. https://www.weusecoins.com/what-is-segwit2x/ https://www.weusecoins.com/what-is-segwit2x/ > such a move would (indeed, will) make major waves in mainstream Joe Pension Fund/ Number Go Up news, who will not be familiar with nor interested in the technical arguments for the fork that is being actively suppressed. there's no weeks-long debate window here. Fair point, given how the markets reacted to this drop in hashrate.

rantwasp 5y ago

that’s not how 51% attacks work. you have the miners but you also have the actual bitcoin nodes that are witnessing the blockchain growth.

bonestamp2 5y ago
Do the witnesses have any weight on the consensus?
- ric2b 5y ago
  Depends on who is using them. If it's a large business/exchange: definitely.

dnautics 5y ago

No, it means that theoretically a 51% is possible but the levels of coordination required to do that and the ability of the bitcoin network to figure it out and ignore it are means that it is unlikely to be effective.

sillysaurusx 5y ago
Plus, the model of BTC as “digital gold” is pretty resilient to 51% attacks: if you don’t touch your coins, no one can take them away from you. (A 51% attack can’t reassign coins.) Of course, the price of BTC might tank in response to the news of a successful 51% attack, but it’ll probably bounce back just like every other event in BTC history.
- paulmd 5y ago
  > if you don’t touch your coins, no one can take them away from you. (A 51% attack can’t reassign coins.) it can't generate new unsigned transactions but it can roll back the network to a state when someone else owned those coins. from a practical perspective there is no difference, you no longer have those coins.
  - fouc 5y ago
    there's a limit to how far it can roll back the network, it probably wouldn't be able to roll back more than a few days..
    - yuvadam 5y ago
      "Rolling back" the network means successfully creating a new longest chain that forks off of an arbitrary old block. The further back in time you to "rollback" to, the less probable it is for the 51% to successfully mine all those needed blocks to get up to par with the existing longest chain. Realistically speaking, this sort of attack is highly improbable.
      - sillysaurusx 5y ago

c54 5y ago

> doesn't this effectively mean China can force a 51% attack by just strongarming the pools that operate there? Yes, in theory. There's a coordination problem to get all the pools to do what you want, but if they (or anyone) could pull that off, then yes. So then what happens? If a single entity controls >50% of the hash power of the network, they could start finding blocks with arbitrary (invalid) transactions and adding them to the chain. The transactions they'd be able to insert are called 'double-spends', which is that wallet A sends its balance to wallet B, and then in the next block sends the same balance (magicked into existence) to some third wallet C. All this can happen. The way the network is supposed to work though, miners don't just commit new blocks and call it a day, the 'full nodes' have to accept the blocks. The miners are more or less offering up solved blocks to the full nodes for their acceptance. As a full node it's trivial to take a look at this double-spend from A->C and say 'nope this doesn't look like a valid block to me, try again'. At this point if enough full nodes do this, the 51% of 'bad' miners are just wasting their time, the true hashrate of the network is now 49% of what it was a moment ago, and the remaining 'good' miners will continue to find blocks. Business as usual. It's a much harder problem to control >51% full nodes, mostly because they're more geographically distributed. This is because they're pretty simple to set up at home and run, without worrying about up front cost of miners, electricity and noise and so on. A bitcoin full node can run on a raspi with a 1tb external hard drive, and a normal consumer internet connection. Another interesting tidbit is that with control of 51% of the miners, this still only means a malicious actor finds a block first 51% of the time... to get the 6 blocks usually required for confirmation (in a row) is a ~3% chance. Not a perfect safeguard, but just some perspective on how hard it is to execute an attack like this. The idea is that the miners are beholden to the full nodes, who are the actual actors in the distributed consensus system. There's some other game theory around why one would want to do obvious double spends after having that much control (why not just sell the bitcoin and take profits?). But there are definitely reasons a state actor would want to try to destroy the network.

dangero 5y ago
inserting invalid transactions is not the double spend issue. If their hash rate is over 50% they can "unconfirm" a confirmed transaction by making a chain with more work on it. This is how the attack plays out in simplified form: Bitcoin is sent to an exchange by the attacker, confirmed, sold and withdrawn for cash. Attacker then makes a chain with more work on it than the chain where the Bitcoin was sent to the exchange. Now as far as Bitcoin network is concerned, the coins were never sent to the exchange, but the exchange already let the user sell the Bitcoin and cash out. So now the exchange has neither the cash or the Bitcoin. The attacker has both.

jabbany 5y ago

In theory yes, but in practice it's rather infeasible to do 51% attacks IRL --- the main constraint is the people not the math. A 51% attack mainly lets you double spend, so you have to: (1) get a lot of coins and spend them on something (2) start mining a fork where that spending didn't happen (but other real transactions continue to happen) --- your hash power will now need to be dedicated to this new fork at the exclusion of the old "real" fork since all future hashes will diverge (3) wait until the thing you paid for with coins can no longer be "taken back", all while continuing efforts on the alternate fork (4) do the 51% attack to swap out the old "real" fork with your new fork where you never bought the thing, thus keeping the thing _and_ the coins you spent on them (5) all this needs to be balanced with the opportunity cost had you just continued mining on the real chain... Such an attack would trigger immediate attention since at (2) the main chain would suddenly see a drop in contributions. Also realistically most things you can purchase with coins (like fiat or other goods/services) can be relatively easily "taken back" unless a considerable amount of time is spent to launder them --- say if you "buy" a car with crypto and reverse the transaction, what you've really done is just stolen a car with a lot of extra steps... Which means your only real viable path is to redeem for some other cryptocurrency... Now given that a big attack just happened on the largest market cap crypto, there's a pretty good chance the other coins are going to take a hit and exchanges might even temporarily limit activities while this situation is "investigated", greatly reducing what you can actually do at the end of the day. This kind of thing is more reasonable on the smaller market cap "meme" coins where you can just exchange them for for a mass-market crypto.

roofwellhams 5y ago
What if this drop in hash is to use on the 2nd chain?
- Thorrez 5y ago
  Well we have a good explanation of the drop: the power outages. Also with 25% hash power you're going to have a hard time doing a 51% attack.
  - 5y ago

emerged 5y ago

I like how the replies are either yes or “yes, but <12 paragraphs of but>”

Zamicol 5y ago

Yes.

sashimi-houdini 5y ago

Yes it does. And if/when a large enough player finds it in their interest to mount such an attack, they might very well do it. See https://www.crypto51.app/ https://www.crypto51.app/ for a (somewhat outdated) price chart. Remember the first law of cryptocurrency though: "in case of emergency, betray your core principles." In the event of such an attack, the rest of the system (miners, developers) may be able to simply declare the "evil" side of the fork "untrue" and hardcode the other fork as true. The fact that this is exactly opposite to Bitcoin's stated goals is no problem at all. See e.g. Ethereum's response to the "DOA Attack" as an example.

usrusr 5y ago

I've been wondering about possible subtle benefits hashrate dominance could bring other than double spending. I'd be surprised if there wasn't some way to benefit nicely from being able to load the dice of which transactions are committed and which have to wait for some minority miner to win the race? Certainly nothing on the scale that it would pay for the hashrate by itself, but that's already covered by regular fees and reward.