Bitcoin's 10000 TH network is extremely vulnerable

[johnyj]

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

1. No one can get 10,000 such mining boxes manufactured in 3 months, and it is very likely in 3 months the difficulty will rise another 10 times

2. Even if someone manufactured 10K such mining boxes, they need a HUGE site to setup those boxes, and 10MW power supply at least, many cooling towers...

3. Combine 1 and 2, anyone who tried to do this will most likely end up with commanding 5% of the network hash power after 3 months with a huge cost

4. Someone might want to manufacture 100K such boxes to stay ahead of the curve, but then the scale of the site will rise to another astonishing level and they might need a small nuclear power plant to power all those ASICs and cooling towers, all these could take years to build

5. Finally, even if someone successfully commanded 60% of hash power, what he actually can do is very limited. He won't be able to affect existing hodlers' coin, just make the future transactions less trustworthy, so once people find out, they will stop transactions and hold their coins and wait for the situation to resolve

[1714983098]

1714983098

[1714983098]

1714983098

[1714983098]

1714983098

[1714983098]

1714983098

[alexeft]

Current network power is ~11500 Th.
To 50% this one would need ~3800 pieces of the latest announced knc miner at ~$10k each, a total cost of ~$40 mil
just for hashing machinery.

If the difficulty continues to rise by 25% each time, in one year the cost to 50% would be ~$8.5 billion.
In 2 years it would be ~$1.79 trillion.

Admittedly, the cost to 50% bitcoin looks too low right now but I think it can be easily explained by the fact
that we are only a few months in the era of asics. Given time, the picture will be totally different.

[LiteCoinGuy]

http://www.coinometrics.com/bitcoin/brix

[Gabi]

here they give really higher 51% attack cost:
http://www.coinometrics.com/bitcoin/brix

That website is epic wrong. Yes i have read their "assumptions" but these definitely does not justify such a high cost.

If the difficulty continues to rise by 25% each time, in one year the cost to 50% would be ~$8.5 billion.
In 2 years it would be ~$1.79 trillion.

Or maybe not. Do you realize that difficulty is rising because much better hardware is being released and NOT because more people invest? We moved from GPU to lower performance ASIC and now moving to higher performance ASIC so difficulty is skyrocketing, but it is exactly like when we moved from cpu to gpu mining. Same investment but much higher difficulty, because the hardware was different.

[suprememilo]

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

1. No one can get 10,000 such mining boxes manufactured in 3 months, and it is very likely in 3 months the difficulty will rise another 10 times

Intel or another semiconductor company could easily do it.

[newguy05]

many companies can easily do it, all you need is a data center and a few thousands blades.  Those banks have dozens of data centers alone just for their high frequency trading, this is like buying some office supply for them. Just need to get the asic chip designed, then crank out the units, install in a data center, turn it on at once. boom...

[johnyj]

here they give really higher 51% attack cost:
http://www.coinometrics.com/bitcoin/brix

That website is epic wrong. Yes i have read their "assumptions" but these definitely does not justify such a high cost.

If the difficulty continues to rise by 25% each time, in one year the cost to 50% would be ~$8.5 billion.
In 2 years it would be ~$1.79 trillion.

Or maybe not. Do you realize that difficulty is rising because much better hardware is being released and NOT because more people invest? We moved from GPU to lower performance ASIC and now moving to higher performance ASIC so difficulty is skyrocketing, but it is exactly like when we moved from cpu to gpu mining. Same investment but much higher difficulty, because the hardware was different.

There are definitely much more miners joining the game

My investment on ASIC mining rigs have increased by 10x comparing to GPU era, but my mining income has shrunk to 1/10 of GPU era. I don't think that everyone else increased their investment by 100x, more likely to be 10x more people joining the mining game with 10x more investment

[runam0k]

many companies can easily do it, all you need is a data center and a few thousands blades.  Those banks have dozens of data centers alone just for their high frequency trading, this is like buying some office supply for them. Just need to get the asic chip designed, then crank out the units, install in a data center, turn it on at once. boom...

Sure, once GS wakes up to the fact Bitcoin is a threat to its business, it will destroy Bitcoin from one of its data centres.

Guess I'll believe it when I see it.

[johnyj]

many companies can easily do it, all you need is a data center and a few thousands blades.  Those banks have dozens of data centers alone just for their high frequency trading, this is like buying some office supply for them. Just need to get the asic chip designed, then crank out the units, install in a data center, turn it on at once. boom...

It will take them years to build such data center, and the power consumption is the biggest issue. Most of today's ASIC device manufacturers' hosting service suffered from power supply and cooling capacity for the site and could not expand their capacity above certain thread

For example that cloud hashing company in iceland, I guess they are only capable of maintaining hundreds of rigs

[e521]

I agree with above, this could be done but it will require too much effort and especially time.
I am not sure how many miners are involved nowadays but if we switch CPUs and GPUs back to BTC I think we can go up a few TH

There are many other ways to destroy bitcoin just with a few words, rather than 51% attack (or 60% if you want that extra power)

[JohnsonRobinson]

...Why would somebody waste say 50 million dollars? No institution can simply do this. Although GS is extremely powerful, they can't just do what they want.

A private person could do it, for example in the future a serious Alt-Coin competitor with interest in destroy one coin. That would be the much more likely scenario of an attack, but still extremely improbable, at least for now.

Have to disagree.  50 million is absolute peanuts to these guys.  They dish this out in bonuses each year.  Fed is printing 84 Billion each month and according the Max Keiser, this is expected to go to over 300 Billion with the Bernanke's successor coming on board...Janet Yellen.  

Also, it's anything but a waste if it wipes out something that threatens their business(es), which are worth much more than that.

Question is, since cryptos are certainly here to stay, what's the next biggest coin that solves this?  Peercoin?

[Lauda]

You're assuming that one can get them all UP in the same time, or before the difficulty adjusts. No.

[Bansheroom]

I was just looking at the bitcoin difficulty chart and realized how small this network really is despite the extreme difficulty increases over the last 2 years.  Lets get to it:

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

We are just an ant...not even ant, a microscopic organism in the financial world.  If bitcoin becomes too big or start hurting those bank's bottomline (transaction fees), all they need to do is sneeze and bitcoin is gone. And there is no regulation/anti-competitive laws to protect bitcoin, it is completely de-regulated at this point.

Heck it doesnt even have to be the banks, even a small hedge fund or single person with some cash can wipe out bitcoin in a single instance.

The other ironic thing is the market cap of bitcoin is around $12billion right now compared to the $20million to destroy it. It shows how far removed from reality everything is, i dont think people really thought this through. If bitcoin can be shorted, someone can just short it, then destroy the network for riskfree win.

The difficulty needs to increase by 1000X at the minimum ($10 billion) to provide some bare minimum security to bitcoin's network

Dont agree with what I said about the current state of bitcoin? then provide your reasons, i love to be proven wrong. But just cant find a fault in my numbers.

No fault in the numbers...perhaps something protects bitcoin you never thought of...

I dont know if 10000 TH are availabe to buy right now, i dont think so.
How much TH in Hardware is produced each day and how much and fast can a single company/person aquire of it withtout accelerating the bitcoin-hype too much?

Anyway if you manage this, i assume the community switches all services to litecoin, you have then to buy the new and expensive Scrypt-ASICS (and perhaps AMD) just to see the community switch to Quark or whatever (still?) rich person/company cant compute.

[Realpra]

This has always been an issue and was much cheaper in the past.

I am sure it has been discussed before.

I suppose it would be possible to implement some sort of trust system for nodes. Devs and pool operators would need to work together.

I suggested a trust filter before.

Possible trust filter algorithm:
1. A full node only accepts a block if ONE of the PAST 50-100 blocks were on its "trusted miner" list.
2. Miners could sign their blocks in different ways without requiring total compliance or a protocol change of any kind.

Result:
If one bad actor 51% attacks the network and prevents all transactions, nodes with trust filter enabled will STILL require some competitors blocks to be allowed leading to the 51% denial of service attack to FAIL.

Similarily rewriting block history far back with your own blocks only would fail.

[dyseac]

If the US government wanted to take bitcoin down, why waste $30mill doing so. They can just declare all aspects of bitcoin usage / trade illegal with severe jailtime. With their influence they could then convince all other western institutionalized countries to do the same thus removing the value of the currency.

Much easier than actually "destroying" it - I mean what's the point of that. People using it for illegal activities will just create another coin. or jump from Bitcoin to Litecoin to Dogecoin etc..

I think they're working out a way to make $30 mill from it rather than spend $30 mill destroying it

besides, how many coins did the FBI confiscate from Silkroad? why would they want to throw that value out the door? !

[drewster]

I was just looking at the bitcoin difficulty chart and realized how small this network really is despite the extreme difficulty increases over the last 2 years.  Lets get to it:

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

We are just an ant...not even ant, a microscopic organism in the financial world.  If bitcoin becomes too big or start hurting those bank's bottomline (transaction fees), all they need to do is sneeze and bitcoin is gone. And there is no regulation/anti-competitive laws to protect bitcoin, it is completely de-regulated at this point.

Heck it doesnt even have to be the banks, even a small hedge fund or single person with some cash can wipe out bitcoin in a single instance.

The other ironic thing is the market cap of bitcoin is around $12billion right now compared to the $20million to destroy it. It shows how far removed from reality everything is, i dont think people really thought this through. If bitcoin can be shorted, someone can just short it, then destroy the network for riskfree win.

The difficulty needs to increase by 1000X at the minimum ($10 billion) to provide some bare minimum security to bitcoin's network

Dont agree with what I said about the current state of bitcoin? then provide your reasons, i love to be proven wrong. But just cant find a fault in my numbers.

I want to make sure this is super clear:

ZZZZZZZZZZZZZZZZZZZZZZZZZZZZzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz..............

[proton]

It seems to me that for any bank or profit-driven entity that can invest significantly in Bitcoin mining, it would be better that Bitcoins appreciate as opposed to being made worthless.

Let's suppose by investing 50000 BTC (~$40M as of Jan 2014), I can get 51% of the total hashing power. This should give me 3600 BTC a day.  Sure, I would have electricity costs, but these should be a fraction, my income should be in the order of say 2000 BTC a day.

I can then recover my investment in a few months (even with accelerated complexity & cost) and the rest would be pure profit in Bitcoins.  In this case, I would prefer that Bitcoin appreciates in value and thus I get great returns with my initial investment.

Of course, it would be a different conclusion when a group (or more likely to be a group governments than corporations) decides first the Bitcoin network should be destroyed, rather than them making profit/loss decisions on the new investment.  But, in that case, we should consider that BTC network would respond to this attack in some fashion as well to protect the worth of network. 
   

[kalus]

I was just looking at the bitcoin difficulty chart and realized how small this network really is despite the extreme difficulty increases over the last 2 years.  Lets get to it:

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

We are just an ant...not even ant, a microscopic organism in the financial world.  If bitcoin becomes too big or start hurting those bank's bottomline (transaction fees), all they need to do is sneeze and bitcoin is gone. And there is no regulation/anti-competitive laws to protect bitcoin, it is completely de-regulated at this point.

Heck it doesnt even have to be the banks, even a small hedge fund or single person with some cash can wipe out bitcoin in a single instance.

The other ironic thing is the market cap of bitcoin is around $12billion right now compared to the $20million to destroy it. It shows how far removed from reality everything is, i dont think people really thought this through. If bitcoin can be shorted, someone can just short it, then destroy the network for riskfree win.

The difficulty needs to increase by 1000X at the minimum ($10 billion) to provide some bare minimum security to bitcoin's network

Dont agree with what I said about the current state of bitcoin? then provide your reasons, i love to be proven wrong. But just cant find a fault in my numbers.

You're analysis is flawed and $20 million estimate is low for a number of reasons:

1.) In order for someone to pull this off, they would need $20 million worth of mining hardware today. If they start now, it will be months before they can bring the hardware online, at which time the difficulty will be much higher. In less than 5 days, the difficulty will go up another 25% so they will need $25 million worth of hardware.

2.) The analysis only covers the manufacturing and development costs, and ignores the cost of labor to set everything up, power, and facilities to run 10,000 TH. The ignored costs are not small.

3.) If someone really could bring 10,000 TH online today for $20,000,000, that setup would earn almost $3,000,000 in bitcoin per day. It doesn't make any sense for someone to spend $20,000,000 to destroy something that would earn them $3,000,000 per day.

4.) If Bitcoin is destroyed, I have little doubt that users will switch to a non SSA-256 coin like Litecoin. Some other crypto that doesn't contain the flaw would take over, so anyone trying to destroy crypto would end up playing whack-a-mole and waste millions of dollars.

5.) If the user was malicious, I suspect they would have millions of dollars worth of liability and end up getting sued into oblivion. Can you imagine how much negative publicity a company would get for even trying to destroy Bitcoin?

------------------cut thread here ---------------------

[newguy05]

I was just looking at the bitcoin difficulty chart and realized how small this network really is despite the extreme difficulty increases over the last 2 years.  Lets get to it:

1) Currently difficulty is at ~10,000 TH
2) Single 1 TH mining box cost ~$1000 to manufacture (reality is much less than 1k, but for easy calculation lets use $1k)
3) To destroy the bitcoin network you need to control 51% of the network.
4) To bring 10,000 TH (assume evildoer wants to be thorough) online it will cost:  10,000 * $1000 = $10million,  add on another $10millon for R&D,  $20million total.
5) Goldman saches makes ~$30million a day in NET revenue (same with most other banks)
6) It will cost a SINGLE bank HALF day's net revenue to wipe bitcoin out - less than their office supply budget.

We are just an ant...not even ant, a microscopic organism in the financial world.  If bitcoin becomes too big or start hurting those bank's bottomline (transaction fees), all they need to do is sneeze and bitcoin is gone. And there is no regulation/anti-competitive laws to protect bitcoin, it is completely de-regulated at this point.

Heck it doesnt even have to be the banks, even a small hedge fund or single person with some cash can wipe out bitcoin in a single instance.

The other ironic thing is the market cap of bitcoin is around $12billion right now compared to the $20million to destroy it. It shows how far removed from reality everything is, i dont think people really thought this through. If bitcoin can be shorted, someone can just short it, then destroy the network for riskfree win.

The difficulty needs to increase by 1000X at the minimum ($10 billion) to provide some bare minimum security to bitcoin's network

Dont agree with what I said about the current state of bitcoin? then provide your reasons, i love to be proven wrong. But just cant find a fault in my numbers.

You're analysis is flawed and $20 million estimate is low for a number of reasons:

1.) In order for someone to pull this off, they would need $20 million worth of mining hardware today. If they start now, it will be months before they can bring the hardware online, at which time the difficulty will be much higher. In less than 5 days, the difficulty will go up another 25% so they will need $25 million worth of hardware.

2.) The analysis only covers the manufacturing and development costs, and ignores the cost of labor to set everything up, power, and facilities to run 10,000 TH. The ignored costs are not small.

3.) If someone really could bring 10,000 TH online today for $20,000,000, that setup would earn almost $3,000,000 in bitcoin per day. It doesn't make any sense for someone to spend $20,000,000 to destroy something that would earn them $3,000,000 per day.

4.) If Bitcoin is destroyed, I have little doubt that users will switch to a non SSA-256 coin like Litecoin. Some other crypto that doesn't contain the flaw would take over, so anyone trying to destroy crypto would end up playing whack-a-mole and waste millions of dollars.

5.) If the user was malicious, I suspect they would have millions of dollars worth of liability and end up getting sued into oblivion. Can you imagine how much negative publicity a company would get for even trying to destroy Bitcoin?

------------------cut thread here ---------------------

There are actually some good responses to find fault in my logic, but the response you quoted is actually one of the worst (no offense to the poster), every single point is just flat out wrong.

Understand spending 20 or 50 million to bring 10-50k TH online is nothing in the real business world, we are not talking BFL or Avalon here. Lookup how much a single sp500 bank makes in revenue in a day, lookup how many hardware/chips apple/Foxconn/tsmc can dump out in a day.  Did those BFL/Avalon scams really conditioned the community to think mass producing asic boxes is difficult?

Understand the corps don't give a shit about mining 3600 or whatever btc a day to make a few million $, lookup how much they make in transaction fees, wire fees, credit card fees. All will be gone if btc fulfill its intended purpose.

Also understand you can replace 1 currency network with another, but to actually have people trust and support the new network, is extremely difficult. See replace gold with copper analogy earlier.

Finally understand there is no liability, implied to otherwise. Btc is not legal tender under any country including us. There is no regulation nor law to prevent company or person to destroy it.

As I said only reason none bothered is because it is not popular enough to affect their bottom line yet. Difficulty needs to catch up - at minimum another 1000x increase before it becomes popular, otherwise there will be trouble.

[lumierre]

It's bound to happen that total mining hardware cost:BTC market cap ratio is going to continue decrease as the rewards from mining get lower. Security will get lower and lower as rewards to mining taper. It's a fact that not most people look at. If Bitcoin does keep a constant  mining hardware cost:BTC market cap ratio until block rewards reach close to 0, the transaction fees will be much much higher than today (Heck, it's already expensive today). That's where Proof-Of-Stake coins such as Peercoin come in. It's quite an innovation not really taken seriously. It's nice to see brand new coded cryptocurrencies realizing this.