Bitcoin Forum

I was wondering if there is any realistic systemic risks if an attacker with enough funds (see, fiat cartels, with their various tools such as futures, printing funds and whatnot) are able to suppress the price to put miners into bankruptcy in order lower the hashrate (thus security of the network) and once decreased enough, launching a 51% attack of sorts or any other malicious action that could be performed upon a weakened network (as miners left due lower prices). We've already seen a decrease in hashrate and many miners forced to quit.

Is this FUD and an non issue because the numbers are still oto high to make it work for the attacker? or is that a real threat that cloud cause a chain reaction of miners leaving?

I've started a somewhat related topic here (https://bitcointalk.org/index.php?topic=5069448.0), so let's see ;)

IMO, most big miners won't leave from one day to another. They have lots of costs to pay not related to electricity (renting the building, labour costs). So as long as they think the Bitcoin price will increase again, and they have the possibility to do it, it makes sense for them to keep on mining if the price is slightly under their "break even price".

So the attacker must be able to lower the price for a longer timeframe than just one or two weeks, and the price must get really low in a short timeframe.

If the attacker has these funds, it's likely he could also start a 51% attack from the start, without having to buy and then sell lots of Bitcoins. There is one slightly more dangerous option: Shorting. But the supply of Bitcoins available for shorting is very likely too small to move the price that significantly to massively drive miners out.

Security is not just about how low attack costs are, incentives should b taken into account as well. Bitcoin price is down like 50% in November but hashrate declined just 25%. It implies a highly de-incentivization of attacks in spite of moderate improvement in the costs of carrying out them. Outcome? Improved security!

It is the beauty of bitcoin and PoW, you have not to worry about security because of its self-regulatory mechanism. The same factor that de-incentivizes mining and decreases hashrate secures the network by making it less temptative for adversaries. On the opposite side, price would never go (and stay enough) that high to make bitcoin insecure because once it approaches that thresholds, users feel unsafe and start dumping their funds and correction happens.

As of 50%+1 short range double spend attack: Never mind! It would be the victim's fault not to wait for enough confirmations when trading multi million dollar assets. I think exchanges and traders have learned their lessons after latest btg double spend which was really embarrassing.

I doubt a 51% attack would be viable (or close to viable) just yet, the cost is simply still too high.

Problem being, where to get the mining hardware? You'd still need current the hashrate and them some to mount an attack.

The currently known large mining hardware producers are unlikely to defect this way, as that would be throwing away a profitable million dollar business.

The hardware that dropped off the network could probably be bought for cheap but that would be (a) an enormous logistical challenge, assuming we're talking about multiple unrelated operations across the globe, (b) hardware that is most likely outdated, given that unprofitable miners will drop off first and (c) still way too little hashing power.

What is left is either (a) buying miners wholesale from mining hardware producers or (b) developing mining hardware from scratch. Both incredibly expensive and both options that would take a lot of preparation time (ie. in case of (a) only a limited quantity of miners are in stock at any given time so you can't buy all at once, in case of (b) research and development and having the miners produced).

In short, the only ones potentially capable of attacking Bitcoin, even considering a continuous decline of hashrate, are the ones that already have skin in the game. At the point that Bitcoin becomes effectively attackable for an outside adversary it's likely not worth attacking anymore anyway.

@aliashraf: Good point, I think we discussed it in another thread. While price/market cap is a debatable indicator, indirectly it influences the profit expectations of an attacker (or better: all potential attackers). Only for destructive attacks (with the goal to "kill" Bitcoin without mattering costs, e.g. started by a government) the incentive is not tied to price.

For these "destructive" attackers, maybe bear markets with shrinking hashrates are an opportunity. It's however not sure if a "destruction by 51% attack" is even possible, as the community can most likely escape with a snapshot followed by a hard fork with PoW algorithm change.

Agree.

I think an attacker having the funds to buy the necessary hardware should also have the funds to develop it straight away and pay a foundry to produce the ASICs. Also, an 51% attack most likely isn't a goal accomplishable in some days or weeks - but the attacker has all the time he wants to prepare it (and buy/develop/rent hardware).

Oh yes that's true. To elaborate, I was thinking about the case of an attacker buying / developing / producing a sufficient amount of mining hardware at rather short notice. That I deem close to impossible, even given generous funding (ie. money can speed things up only within a limited margin as the law of diminishing returns kicks in).

Time for preparation may be plenty, but once the preparations are done the clock is ticking -- an attacker only has a limited timeframe before their hardware becomes obsolete.


Come to think of it, I believe there's a flaw in trying to suppress Bitcoin's price to make 51% attacks easier -- you'd have to buy in first, which would likely prop up the price and consequently the overall hashrate. I might underestimate OTC markets and the impact of derivatives though.

It has been 10 years and hashrate has not been always that high and no "destructive" attacker with an irrational agenda has ever showed up.

And it is the point: bitcoin is not designed to resist crazy players ready to burn everything including his assets to ashes but let's take a look at such a hypothetical scenario:
A crazy person or entity ready to burn like billions of dollars to destroy bitcoin for fun or for some political agenda: e.g. Trump fails to making america great again by destroying Iran because they use bitcoin as a detour to buy medicine or food or cellphones so, he decides to ruin bitcoin by spending 1-2 billion dollars (probably he blackmails Bin Salman to cover the budget) now what happens?

Trump buys like 10,000,000 s9s (from nowhere) and plugs them to a huge 15,000 MW power plant. what are the costs? No matter, Bin Laden Salman would pay.  :D

Now what? Idiots who are running the attack on behalf of Trump, avoid confirming transactions and generate empty bloks while rejecting occasional blocks produced by loyal miners, bitcoin looks to be dead, well it is not, yet!

1- Attackers won't be able to rewrite the blockchain for more than a specific length because a work tens of thousands greater than what they can afford is already accumulated in the history. So, history won't be rolled back very deep.

2- It is a matter of few hours for the jump in hashrate and malicious actions to be discovered and people start considering to fight back.

3- Within 24 hours loyal miners stop operations and Trump crew slow down too, it helps them keeping costs lower. Bad news but definitively not the end as opportunity costs are still high besides human resources, ....

3- The first reaction would be using checkpoints in the code to preserve the history up to the attack start time.

4- The next step would be forking to another algorithm making Trump-Bin Salman investment totally void.

Now they have to recruit new forces and continue chasing us in case Trump gets re-elected and is still pissed off about his failure. We are screwed, but no worries, Bin Salman is screwed too  ;D

OP, realistically? I believe there is none. The "attackers" would be in a better situation if they mined honestly and supported the network, or invest their money in Bitcoin, instead of spending it to "kill" Bitcoin.

If Bitcoin was really dying, then no one needs to do anything but watch.

            Selling BTC on centralized Fiat exchanges does not lead to fair price and transparency. Just take a look at precious metals.

I do see the following potential danger in a (significantly) decreasing hashrate:

• Larger miners/pools stop mining due to lower prices
• Mining hardware is still around, also if not used to mine Bitcoin
• Assume a large miner that has had enough of it and wants an exit
• Due to decreased hashrate it is possible that this miner alone (or e.g. 2 miners combined) has enough hashing power for a 51% attack
• With a successful 51% Attack the miner can reverse TXs
• The miner sells all his bitcoin to some investor
• Reverses the TX
• Sells again to another investor
• Reverses the TX
• ...

Of course, the iterated selling will be for decreasing prices, but since the miner wants to exit he couldn't care less. At the end the miner will have made the best possible exit while trust in Bitcoin will be gone.

@aliashraf: There's one variant of the "destructive attack" that could worry me:

If Bitcoin changed to an ASIC-unfriendly algorithm (like ProgPoW, where ASICs are only two times more efficient than CPU/GPUs, if I remember correctly), then the "destructive attacker" could simply buy standard computing equipment, if he has the funds to buy the necessary hashrate. The "PoW change" option then would become extremely difficult, as if we changed back to an ASIC-friendly algorithm, at the start there wouldn't be any ASICs available - so the attacker had still time to continue his attack with standard computing equipment.

So one could even argue that for Bitcoin it's better not to change to an ASIC-unfriendly algorithm, but to an algorithm suitable for differently constructed ASICs (Scrypt, for example?). I'm not totally sure about that, though.

@GaPR: I don't really understand your answer to my post ...

I realized that on the positive side: most competent miners plan ahead, if not all of them, so they should be ready to mine at a loss for a long time, speculating that the price will go up in the future. A lot of them have done this in the past with great results, and there's no reason bitcoin will not go to ATH again since the fundamentals are the same as always, this is just fiat price manipulation as we have seen every 3 years + FOMO.

Also ASIC prices should go down if many miners start quitting as they would be panicking wanting to get rid of all their gear as they would think it would become worthless, so someone else out there would see an opportunity to get these cheap ASICs and start mining.

I think the risk would only be if there was a massive and sudden drop of hashrate, but we have so much hashrate that the losses we've seen don't matter that much.

@d5000
Your point about 'destructive attacker" being capable of buying/hiring commodity computing devices is correct but ProgPow for instance is gpu friendly and gpus are rare enough to make it very hard for the attacker and interestingly they are far more distributed than ASICs.

I don't think there could be any solution for such an attack other than making it harder and more expensive to be carried out.

Now let's rethink our hypothetical destructive attack a bit more:
As of now, we have succeeded to make it really expensive for the attacker and still there is a chance for him to take the bid and commit such an attack, but why? I would suggest that we are speaking of a situation in which we have bitcoin as a serious financial device with secondary socio economic and political side effects like jeopardizing establishment in financial sector like central banks whatsoever.

Actually I think it is a good assumption to make that we won't face such an attack unless bitcoin is adopted like 100 times more. But more adoption is not feasible without higher prices and it implies a network with an order of magnitude higher hash power, hence orders of magnitude higher attack costs. I'm less concerned  as far as it is about destructive/irrational attacks to bitcoin, if not being 100% convinced about my security.

But in a realistic scenario, how can Bitcoin hard fork to a new POW algorithm, with consensus, if what it will do is amputate the network of full nodes and hashing power? It will reduce the network's security by a lot.

The proposal of the idea itself is already stupid, unless the mining cartel started to censor transactions.

I just don't see a reduction, here's why.

I get daily bs from bitmain about $100 s-9's, they can't even give them away,

So people buy them plug them in, and use them as space-heaters, now as normal 'miners for profit' pull the plug, these people who don't care, don't even notice that their s-9 is calling home and finding btc for bitmain, and maybe even splitting a nano-profit for our guy in the cold chinese flat. Ignoring the problem that an s-9 sounds like a jet engine, but I'm sure somebody will resolve this problem.

Thus yes, miners are dropping out, but these home heater miners will continue to be deployed, because people have to pay for space-heat, might as well toss in a lottery ticket and the chinese love this shit,

OK, compeche?

Impossible, people who say this are SW people who don't understand HW, the HW people have already said, "No matter how difficult you make your algo or how much memory it uses, we will emulate that algo on an ASIC".

Thus at this point making an ASIC resistant algo "POW" is just another example of retardation in the bitcoin (SW) peoples narrative world-view.

There is NO algo on earth that BITMAIN can't put on an ASIC.

I seem to remember proposals of doing a gradual shift from SHA-256 to whatever hash Bitcoin would use next, precisely as to avoid (or at least alleviate) this issue, however I can't find the respective threads right now. IIRC this would include a transitional period of hybrid mining. AFAIK these proposals haven't been discussed in earnest though and I'm not quite sure about their viability.



I doubt that "HW people" would say that, given that you emulate hardware using software, not the other way round ::)

it was me who proposed this in few contexts, not in a dedicated topic tho. For instance, discussing Chinese pools/ASIC manufacturers 50%+1 threats in a topic started by myself:

Unfortunately, I didn't receive enough support for it. Apparently people are busy trading bitcoin rather than developing it, the true reason behind bitcoin recent fall, imo.

this is what I have tweeted about the future of crypto-currencies and their immediate need to CHANGE:

"it is not the most intellectual of the species that survives; it is not the strongest that survives; but the species that survives is the one that is able best to adapt and adjust to the changing environment in which it finds itself -- Professor Darwin"



P.S.:

sometimes that people talk about the huge hash power of bitcoin and call it STRONG, I automatically recall this quote from Professor Darwin..

Well, the attacker has a lot of time to buy the equipment. While the time slot is not infinitely long because of new hardware generations, I guess GPUs are a tool with massive enough production to give the attacker a real possibility to succeed in buying enough units.

But I agree with your other statement:


... and so I also agree that even a destructive attack is very unlikely. It's not more likely than a massive crash in fiat currencies or other assets. Yep, I agree that Bitcoin is secure - and the recent price crash hasn't changed anything about that.

Regarding a hypothetical hash algorithm change: There are two problems in an emergency situation (colluding SHA256-Miners attacking the network continuously, and we need another algo to stop it):
1) the change needs to be carried out fastly (we don't have the time for a smooth transition, as otherwise miners would have, probably, ways to interfere in the transition period)
2) the attackers could have prepared for the change. So maybe they're already hoarding GPUs to attack the nascent (and thus relatively unsecure) Asic-unfriendly coin.
If this situation emerges, Bitcoin could have a hard time - I don't see convincing solutions for this scenario. Maybe the Bitcoin devs could choose a random algorithm from a group of candidates, to prevent a fast GPU attack? (I think I remember a discussion about that topic in another thread related to the Segwit change).

New generations of gpus take too long to be introduced and commercially justifiable for mining, typically they start with very high prices, more attractive for game enthusiasts. The time window would be large enough to wait for other socioeconomic and political factors to help bitcoin for surviving brutal actions.

1) It is why I'm proposing a hybrid dual algorithm fork, right now. Please refer to my previous post above thread.

2) Being prepared for both ASIC and GPU dominance is more costly, it is not an ultimate solution, I do agree but it de-incentivizes such an attack even more. Although governments have huge resources available they need to justify their actions somehow, somewhen.

But what made you believe that such a proposal will not be contentious, or be supported by the miners? Like franky1 said, miners are also part of the Bitcoin community.


Because it would be contentious.


Whose fault is it? You must be trolling.

By miners you mean pools, right? And I got a solution for getting rid of pools.  ;)

Whose fault is what? Ignoring ever changing environment and the threats for bitcoin?
I guess people who resist change because they have "invested" too much and spread FUD about serious improvements in bitcoin, no matter how critical would be the situation, they think 'short' about bitcoin and guess what, they are losing even more because of their stupidity.

By the way, how is it possible to call the same post as an act of trolling when it is merited by two respected forum member besides its absolute relevance to the topic and the discussion. Your vanity post looks to be more adequate for such a label!

I don't know if I understand you correctly, but doesn't that confirm the danger of a malicious GPU attack? If nobody mines with the newest GPUs, then the attacker has time to wait until GPUs are cheap, start to buy them, and start an attack some months later. That's why in an anterior post I wrote that it's maybe better to change to an alternative Asic-friendly algorithm instead of a GPU/CPU algo.

Yes, I read it. The problem I pointed out is exclusive to short-term "emergency" changes. Unfortunately, after a gradual transition to a new algorithm, the risk for a need for an "emergency exit to another algo" does not disappear completely. What disappears, however, is the dependency on major ASIC manufacturers - and that could be positive, as it's not entirely clear if big players like Bitmain will stay honest, and if not, Antbleed may be a danger.

For now, I (cautiously) support your idea of a gradual change to a ProgPoW-like algo, but it needs more research, I think.

In a "fast emergency algo change" scenario, the problem is the first phase after the hardfork when hashrate is low - as GPUs are standard hardware their power could even be rented more easily than current "cloud mining" power.  I guess that for an entity having attacked the (ASIC-secured) Bitcoin successfully in a "destructive" fashion, the additional cost for the attacker in this first GPU phase should be relatively low.

@d5000
Here is my thoughts about a destructive irrational attack against bitcoin:
1- A destructive attack against bitcoin is unlikely to happen but is more possible when hashrate drops significantly because of price falling deep. it is theoretically possible for governments (US, China) to buy cheap retired ASICS and attack the network just to prove some political agenda.

2- If hypothetically bitcoin was gpu mined, a no more profitable mining situation is more tolerable for gpu owners as they are still usable for game, rendering, ML, ... applications. Dump happens but not to the extents that ASICs would be dumped.

3- GPUs have more uniform distribution across the planet and have a lot more inertia than ASICS. It is much harder to have a secret GPU  manufacturing and preparation phase unlike ASICs because they use high quality RAM which is a rare global resource actively monitored by heterogeneous stakeholders all around the world. It is very unlikely for an adversary to compete with the rest of the planet in terms of equipment.

4- Technological improvements does not linearly increase gpu hashing performance because they need to improve memory unit as well and it is already pushed to its edges of optimization and is not expected to experience sharp developments in foreseeable future.

5- In my proposal, we will never eliminate SHA256 instead we gradually decrease its weight in an scheduled time table, when it comes to emergency, it would be easy to postpone or expedite the process properly.

Will the "solution" get consensus?


But you are talking under the hypothesis that ProgPow will be ASIC-resistant. Reality has already shown that companies like Bitmain will make ASICs for any mining algorithm if a cryptocurrency is succesful enough because they are good at it, and it is profitable.

Plus also under the belief that it will get consensus. Don't you know how dangerous contentious hard forks are? Have we not learned from Ethereum.


Which "people" were in your mind when you said, "Apparently people are busy trading bitcoin rather than developing it, the true reason behind bitcoin recent fall, imo." The developers?

Do you believe those "people" really caused the crash?

@Wind_FURY

I'm pretty much sure, you are derailing this topic. Bitcoin hashrate is dropping sharply and you are worried about your beloved devs not to be criticized even indirectly. Take it easy and move on.

FYI: my criticism target was mainly the investment sector, the people who are concerned about and interested in price to "skyrocket" and do not understand for this to happen we need bitcoin to improve.

Sorry, but it wasn't my intention.


I am confident that Bitcoin's difficulty adjustment feature will work as Satoshi intended it to be. We might also be experiencing longer intervals between blocks now, but it will be back to the normal average of 10 minutes after some time.


You can criticize, but they are the rightful stewards of the network because they have Bitcoin's interest at heart in my opinion. They will not be pressured to make stupid changes to the protocol that might break it, or cause a chain-split, simply because someone wants to change it.


Will you criticize them in a bull market? 8)

I have relaized this is nonsense and has happened after every big crash. Please refer to this:


https://twitter.com/matt_odell/status/1006968132001517584

Then months later we have this:

https://coingape.com/bitcoin-price-predictions-got-wilder-predicts-bitcoin-will-die/

The guy on that tweet predicted it.

There's also a good article that debunks the mining downward spiral collapse:

https://coinjournal.net/debunking-the-bitcoin-death-spiral-theory/amp/?__twitter_impression=true

It couldn't happen due the incentives breaking, I only see a deliberate attack by TPTB looking for a weaker network moment with decreased hashrate. However if they haven't done it years ago, I don't see how can they pull it anymore, the amount of hashrate is ridiculously high.

Even if after massive expenses they managed to break something, they would just get forked off as no one would support an attacker's blockchain.