[2019-01-09]Bitcoin And Ethereum Blockchains Are Not Protected From 51% Attack

[Alex077]

Bitcoin And Ethereum Blockchains Are Not Protected From 51% Attack

On January 5, 2019, the Ethereum Classic (ETC) cryptocurrency blockchain was under 51% attack. Could the similar incidents happen in the future again?  Leading experts in the blockchain industry have expressed their opinion on this issue.

Blockchains Are Not Protected From 51% Attacks
The president of Blockchain at Columbia, Nir Kabessa, believes the main goal of 51% attack is an attack on blockchain of popular cryptocurrencies. However, an attack on Bitcoin and Ethereum, will require large capital investments and coordination of several large mining pools actions. Hashrate of ETC is about 22 times less than the Ethereum network one (8 TH/s versus 177 TH/s). That's why the 51% attack on the Ethereum Classic blockchain costs 22 times less than the attack on the Ethereum network. According to Crypto51, the 51% attack on ETC blockchain costs about $ 4.722, and on ETH network - $ 102.509...https://en.bit.news/bitcoin-ethereum-blockchains-not-protected-51-attack/

[1Referee]

It's very unlikely that something like this will happen with Bitcoin just for the sake of double spending, and there isn't enough hashrate available to set up such an attack. Miners, regardless of how much we dislike them, will do everything they can to keep the main network operating as healthy as possible.

There is more money to be made over a longer period of time by running normally, than to attempt a double spend, which will likely fail anyway.

On top of that, the 'hashwar' that we have seen between ABC and SV also shows how pools are ready to protect a network by pooling their resources together and keep malicious entities outside. It will happen with Bitcoin as well, which means that an attack from whatever unknown entity is even less likely.

[deisik]

It's very unlikely that something like this will happen with Bitcoin just for the sake of double spending, and there isn't enough hashrate available to set up such an attack. Miners, regardless of how much we dislike them, will do everything they can to keep the main network operating as healthy as possible

But the possibility is still there

And while for most governments such an attack is definitely prohibitively expensive, that may not be the case for governments of leading world powers such as China or US. I wouldn't be surprised to find out one day that they actually considered and estimated the costs of carrying out this initiative. Right now it is unlikely but what if Bitcoin becomes a real danger to their financial hegemony?

There should be ways of preemptive countermeasures which are to be explored in advance. Ultimately, POW is not very energy efficient, let's face it, so on-chain scaling goes straight out the window. LN solves this issue, but it also makes a lot of hash power redundant in the future. The latter basically means that the possibility of a 51% attack rises with the hashes diminishing down the road

[KingScorpio]

It's very unlikely that something like this will happen with Bitcoin just for the sake of double spending, and there isn't enough hashrate available to set up such an attack. Miners, regardless of how much we dislike them, will do everything they can to keep the main network operating as healthy as possible

But the possibility is still there

And while for most governments such an attack is definitely prohibitively expensive, that may not be the case for governments of leading world powers such as China or US. I wouldn't be surprised to find out one day that they actually considered and estimated the costs of carrying out this initiative. Right now it is unlikely but what if Bitcoin becomes a real danger to their financial hegemony?

There should be ways of preemptive countermeasures which are to be explored in advance. Ultimately, POW is not very energy efficient, let's face it, so on-chain scaling goes straight out the window. LN solves this issue, but it also makes a lot of hash power redundant in the future. The latter basically means that the possibility of a 51 attack rises with the hashes diminishing down the road

well thats all old news, the hasrate is collapsing due to price collapse, so its getting more and more likely

leased pos will be future like waves.

ps: your avatar is remarkable how can i get something like that

[figmentofmyass]

there's a big difference between "ethereum classic" and "the most secure network in the world (bitcoin)". the incentives and costs to attack the network are completely different.

It's very unlikely that something like this will happen with Bitcoin just for the sake of double spending, and there isn't enough hashrate available to set up such an attack. Miners, regardless of how much we dislike them, will do everything they can to keep the main network operating as healthy as possible

But the possibility is still there

And while for most governments such an attack is definitely prohibitively expensive, that may not be the case for governments of leading world powers such as China or US. I wouldn't be surprised to find out one day that they actually considered and estimated the costs of carrying out this initiative. Right now it is unlikely but what if Bitcoin becomes a real danger to their financial hegemony?

that's the only type of scenario where such an attack makes sense---where incentives external to the protocol become large enough to attack the system in spite of the costs. attacking bitcoin is nothing like attacking ETC due to the capital required. just sourcing enough hardware for the attack might be nearly impossible for anyone but the largest of state actors.

[1Referee]

Ultimately, POW is not very energy efficient, let's face it, so on-chain scaling goes straight out the window. LN solves this issue, but it also makes a lot of hash power redundant in the future. The latter basically means that the possibility of a 51 attack rises with the hashes diminishing down the road

How does it make a lot of hash power redundant? Miners will always bank on the on-chain transaction fees, plus the transaction fees they generate from merge mining side chains, which is something that will become more and more of a dominant aspect in the forthcoming years.

In the end, I strongly believe that miners will make sure they run on energy they generate themselves rather than buying it in bulk from whatever energy supplier. In other words, the energy cost will continue to shrink per miner, which is something that's already happening to a certain degree. It's how markets work, every participant will do everything possible to make sure that operational cost shrinks and profit margins grow.

[deisik]

Ultimately, POW is not very energy efficient, let's face it, so on-chain scaling goes straight out the window. LN solves this issue, but it also makes a lot of hash power redundant in the future. The latter basically means that the possibility of a 51 attack rises with the hashes diminishing down the road

How does it make a lot of hash power redundant? Miners will always bank on the on-chain transaction fees, plus the transaction fees they generate from merge mining side chains, which is something that will become more and more of a dominant aspect in the forthcoming years

This is only an assumption

So let's assume that the majority of transactions goes through LN payment channels. This is a valid assumption if we accept that 1) Bitcoin's adoption expands dramatically in the coming years, and 2) on-chain scaling is a dead-end (which is kind of obvious). In this case, it is possible that the number of on-chain transactions decreases despite Bitcoin's expansion, which, coupled with halved and then halved again reward, makes miners leave

But even if they don't leave, the very idea of POW reminds me of the on-chain scaling itself. Just imagine that the world economy depends on a bunch of miners. This is outright insane and it alone makes the whole POW concept inadequate and rudimentary in and of itself - as much as on-chain scaling is

[BitHodler]

With lots of money, PoW (and all decentralized consensus method) isn't safe from 51% attack or other method of attack.

Unless the attacker work together with pool & miner, it's hard to launch 51% attack since obtain ASIC/GPU in huge number, optimal location and huge electricity amount are hard task even for many government.

I'm sure that there are enough rogue Ethereum miners out there with gear collecting dust due to the lower prices. What fun things can you do with all that unused gear? Right, attack smaller chains.

I hope for smaller chains that Ethereum's POS move will take a while, because this is legit a threat to everything they stand for, and of course the exchanges themselves, because they have the most to lose.

The most logical step for exchanges is to increase the number of confirmations needed for deposits with orders of magnitudes, possibly thousands, which some exchange do right now.

[squatter]

So let's assume that the majority of transactions goes through LN payment channels. This is a valid assumption if we accept that 1) Bitcoin's adoption expands dramatically in the coming years, and 2) on-chain scaling is a dead-end (which is kind of obvious). In this case, it is possible that the number of on-chain transactions decreases despite Bitcoin's expansion, which, coupled with halved and then halved again reward, makes miners leave

Lightning requires commitment and settlement transactions -- these occur on-chain. If Lightning sees exponential adoption, it will greatly increase on-chain throughput. There's no way around that. That's why many people concede that even if we offload most throughput onto LN, we'll still need to increase the block size limit in the future.

[hatshepsut93]

There should be ways of preemptive countermeasures which are to be explored in advance. Ultimately, POW is not very energy efficient, let's face it, so on-chain scaling goes straight out the window. LN solves this issue, but it also makes a lot of hash power redundant in the future. The latter basically means that the possibility of a 51% attack rises with the hashes diminishing down the road

Consensus algorithms and scaling have absolutely nothing to do with each other.

LN doesn't make PoW redundant, it's a second layer that is based on the first layer. Onchain transactions will become more and more expensive so they will be mostly used for high-value transactions. The blocks will be full and miners will be getting their fees, and LN wouldn't hurt their profits in any way.

[squatter]

LN doesn't make PoW redundant, it's a second layer that is based on the first layer. Onchain transactions will become more and more expensive so they will be mostly used for high-value transactions. The blocks will be full and miners will be getting their fees, and LN wouldn't hurt their profits in any way.

Not only that, but it should increase overall on-chain throughput. LN will theoretically enable lots of transactions that were previously cost-prohibitive. The associated opening and settlement transactions are therefore a net gain for miners.

More importantly, the security of every LN transaction ultimately depends on Bitcoin's mining security. The blockchain arbitrates cases of attempted fraud and theft on the LN network. You can't have LN without Bitcoin miners.

[deisik]

So let's assume that the majority of transactions goes through LN payment channels. This is a valid assumption if we accept that 1) Bitcoin's adoption expands dramatically in the coming years, and 2) on-chain scaling is a dead-end (which is kind of obvious). In this case, it is possible that the number of on-chain transactions decreases despite Bitcoin's expansion, which, coupled with halved and then halved again reward, makes miners leave

Lightning requires commitment and settlement transactions -- these occur on-chain. If Lightning sees exponential adoption, it will greatly increase on-chain throughput. There's no way around that. That's why many people concede that even if we offload most throughput onto LN, we'll still need to increase the block size limit in the future.

There is no logic behind your claims

No one will be creating two transactions to make only one just because of LN (other than testing waters, of course). It should be abundantly obvious that payment channels will be opened to pass through a lot more than just 2 transactions. If it is more like 20 transactions (a completely arbitrary number), the number of on-chain transactions is set to decrease dramatically by a factor of 10 (if we assume that all transactions go through LN channels)

In real life it will be more like thousands of transactions per channel, otherwise LN doesn't make a lot of sense. This is what you can't get around provided LN kicks off for real. LN will kill mining as we know it, and that's a good thing in fact as it is archaic, barbaric and totally out of tune with what Bitcoin aims at, i.e. becoming a world money - reliable, efficient and secure

[squatter]

So let's assume that the majority of transactions goes through LN payment channels. This is a valid assumption if we accept that 1) Bitcoin's adoption expands dramatically in the coming years, and 2) on-chain scaling is a dead-end (which is kind of obvious). In this case, it is possible that the number of on-chain transactions decreases despite Bitcoin's expansion, which, coupled with halved and then halved again reward, makes miners leave

Lightning requires commitment and settlement transactions -- these occur on-chain. If Lightning sees exponential adoption, it will greatly increase on-chain throughput. There's no way around that. That's why many people concede that even if we offload most throughput onto LN, we'll still need to increase the block size limit in the future.

There is no logic behind your claims

No one will be creating two transactions to make only one just because of LN (other than testing waters, of course). It should be abundantly obvious that payment channels will be opened to pass through a lot more than just 2 transactions.

How does that contradict anything I said? Obviously people will use LN to make multiple transactions per on-chain commitment transaction. That's the whole point. But anyone who wants to use LN needs to make a commitment transaction, and likely multiple to have decent liquidity, plus periodic funding/re-balancing transactions. The more people that use LN, the more on-chain transactions are required. LN can still provide exponential scale, but you can't ignore that it's still built on top of on-chain throughput.

If it is more like 20 transactions (a completely arbitrary number), the number of on-chain transactions is set to decrease dramatically by a factor of 10 (if we assume that all transactions go through LN channels)

That's a weak assumption. First of all, using LN requires multiple security tradeoffs. People won't stop using Bitcoin just because LN is cheaper. If that were true, altcoins with cheaper fees would have far more usage than they do. Look at Segwit adoption: Users don't even sacrifice any security to get Segwit's cheaper fees and yet adoption still remains below 50%. Secondly, you ignored the argument that LN enables transactions that were previously cost-prohibitive. That means transactions that previously never would have used Bitcoin (micropayments) will now add to on-chain throughput via commitment and settlement transactions. That's a net gain for miners.

[deisik]

So let's assume that the majority of transactions goes through LN payment channels. This is a valid assumption if we accept that 1) Bitcoin's adoption expands dramatically in the coming years, and 2) on-chain scaling is a dead-end (which is kind of obvious). In this case, it is possible that the number of on-chain transactions decreases despite Bitcoin's expansion, which, coupled with halved and then halved again reward, makes miners leave

Lightning requires commitment and settlement transactions -- these occur on-chain. If Lightning sees exponential adoption, it will greatly increase on-chain throughput. There's no way around that. That's why many people concede that even if we offload most throughput onto LN, we'll still need to increase the block size limit in the future.

There is no logic behind your claims

No one will be creating two transactions to make only one just because of LN (other than testing waters, of course). It should be abundantly obvious that payment channels will be opened to pass through a lot more than just 2 transactions.

How does that contradict anything I said? Obviously people will use LN to make multiple transactions per on-chain commitment transaction. That's the whole point. But anyone who wants to use LN needs to make a commitment transaction, and likely multiple to have decent liquidity, plus periodic funding/re-balancing transactions. The more people that use LN, the more on-chain transactions are required. LN can still provide exponential scale, but you can't ignore that it's still built on top of on-chain throughput

I'm curious whether you really don't see where your reasoning fails

Let's assume there're 250k transactions daily in Bitcoin and all of them get written straight on the blockchain. Now LN starts to be used, and the more people use it, the more on-chain transactions it requires (as you correctly suggest). But as we already assumed (and this is a valid assumption), people are using LN as designed, i.e. to make transactions. If there are 25 transactions per channel on average (for simplicity's sake), the number of transactions opening and closing payment channels will ultimately be 250 / 25 x 2 = 20k transactions. And that's what's left for miners to confirm as all other transactions are now in LN

If it is more like 20 transactions (a completely arbitrary number), the number of on-chain transactions is set to decrease dramatically by a factor of 10 (if we assume that all transactions go through LN channels)

That's a weak assumption. First of all, using LN requires multiple security tradeoffs. People won't stop using Bitcoin just because LN is cheaper. If that were true, altcoins with cheaper fees would have far more usage than they do. Look at Segwit adoption: Users don't even sacrifice any security to get Segwit's cheaper fees and yet adoption still remains below 50%. Secondly, you ignored the argument that LN enables transactions that were previously cost-prohibitive. That means transactions that previously never would have used Bitcoin (micropayments) will now add to on-chain throughput via commitment and settlement transactions. That's a net gain for miners.

Our basic assumption is that LN actually makes sense. If you accept this assumption, then there's only one conclusion you can arrive at (see above)

[squatter]

I'm curious whether you really don't see where your reasoning fails

Let's assume there're 250k transactions daily in Bitcoin and all of them get written straight on the blockchain. Now LN starts to be used, and the more people use it, the more on-chain transactions it requires (as you correctly suggest). But as we already assumed (and this is a valid assumption), people are using LN as designed, i.e. to make transactions. If there are 25 transactions per channel on average (for simplicity's sake), the number of transactions opening and closing payment channels will ultimately be 250 / 25 x 2 = 20k transactions. And that's what's left for miners to confirm as all other transactions are now in LN

No, not at all. You're making multiple horrible assumptions that have no basis in reality.

You're assuming that Bitcoin's total transactions are a static number which LN is cannibalizing. That's completely nonsensical. There are multiple arguments made above as to why it makes no sense. You continue to ignore them in favor of repeating dumb claims. You're essentially saying that all future throughput will be channeled through LN. You haven't done any game theory analysis as to why that would be. Based on the fact that "cheap fees" haven't driven Bitcoin's users to altcoins -- not to mention the additional security tradeoffs of LN like having to keep private keys online -- it's very clear that LN transactions are not a replacement for Bitcoin transactions. You're just assuming they are because you've made no attempt to analyze the incentives or security model.

[deisik]

I'm curious whether you really don't see where your reasoning fails

Let's assume there're 250k transactions daily in Bitcoin and all of them get written straight on the blockchain. Now LN starts to be used, and the more people use it, the more on-chain transactions it requires (as you correctly suggest). But as we already assumed (and this is a valid assumption), people are using LN as designed, i.e. to make transactions. If there are 25 transactions per channel on average (for simplicity's sake), the number of transactions opening and closing payment channels will ultimately be 250 / 25 x 2 = 20k transactions. And that's what's left for miners to confirm as all other transactions are now in LN

No, not at all. You're making multiple horrible assumptions that have no basis in reality.

You're assuming that Bitcoin's total transactions are a static number which LN is cannibalizing. That's completely nonsensical. There are multiple arguments made above as to why it makes no sense. You continue to ignore them in favor of repeating dumb claims. You're essentially saying that all future throughput will be channeled through LN. You haven't done any game theory analysis as to why that would be. Based on the fact that "cheap fees" haven't driven Bitcoin's users to altcoins -- not to mention the additional security tradeoffs of LN like having to keep private keys online -- it's very clear that LN transactions are not a replacement for Bitcoin transactions. You're just assuming they are because you've made no attempt to analyze the incentives or security model.

I think I explained my point abundantly clear

So far the number of Bitcoin transactions is not showing a tendency for an exponential growth (or just consistent growth, for the record). That pretty much means that we can consider this number constant for the time being. Your logic is like that of big blockers - let's make blocks bigger and then pretend they are filled. It doesn't work like this in real life. I don't care about future growth, it may or may not happen, but at this point it doesn't
 
If LN expansion will be outrunning the growth in the total number of transactions, miners will be suffering. If you disagree with this assumption, then the whole argument makes no sense as it is essentially equal to saying that LN is of no use. And while we are at it, stop throwing phrases like "game theory analysis", "additional security tradeoffs", "security model" and their likes as you are producing irrelevant noise