[If tx limit is removed] Disturbingly low future difficulty equilibrium

[cunicula]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

[ttk2]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

No single company has the resources to kill off Bitcoin, a government with a few hundred million could build a card manufacturing plant, and pump out the cards in the needed volume. But that would take years of preparation. Even if they started today no entity on earth could take over the block chain in less than 6 months.

[kjj]

You need to think this scenario through a bit more.  Ask yourself what circumstances must occur for a node to be unable to rejoin the world after a network split.  I'll give you two hints.  For the network to get stuck, there needs to be mining power on both sides of the split.  And the duration necessary for the split to become permanent is directly related to the split ratio in a way that makes an honest split of sufficient size and duration to be extremely unlikely.

I need to research this more, yes. But what I had in mind is a targeted attack against a particular node, trying to isolate it from the network and feeding it blocks distinct from those known to the rest of the network. I think it's possible to do that, and while I don't know if such an attack can be lucrative, the victim node can never recover if it adheres to your protocol rigorously.

But if it turns out to be a viable means for increasing security of large transfers (for which a day of waiting would be acceptable) then by all means, proof of stake won't be needed.

We don't care about one node, we care about the network.  If one node gets screwed, the operator of that one node can just delete the block chain and start over.  Or at least just the bad blocks.  And a node can protect itself by having lots of connections.

[cunicula]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

No single company has the resources to kill off Bitcoin, a government with a few hundred million could build a card manufacturing plant, and pump out the cards in the needed volume. But that would take years of preparation. Even if they started today no entity on earth could take over the block chain in less than 6 months.

This is ridiculous. Ebay owns paypal and has a market cap of 43 billion. That is about 4000 times as much money as the appromiximate 10 million required to disrupt block processing. If they want to they could stop bitcoin within a month. It is chump change to them. If they took bitcoin as a serious threat they would probably have been killed it off already.

[ttk2]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

No single company has the resources to kill off Bitcoin, a government with a few hundred million could build a card manufacturing plant, and pump out the cards in the needed volume. But that would take years of preparation. Even if they started today no entity on earth could take over the block chain in less than 6 months.

This is ridiculous. Ebay owns paypal and has a market cap of 43 billion. That is about 4000 times as much money as the appromiximate 10 million required to disrupt block processing. If they want to they could stop bitcoin within a month. It is chump change to them. If they took bitcoin as a serious threat they would probably have been killed it off already.

It would be a lot more than that

Cost is not the limiting factor when it comes to preforming a 51% attack, taking over the block chain is cheap, the problem is getting the cards. As it is the small mining community has already made a major impact in the supply of high end ATI cards, it would take months for any person to get their hands on enough cards to preform a serious attack. Any attempt to acquire such a large number of cards so quickly would throw up red flags all over the place and give the nation under attack ample time to prepare. Any nation attacking Bitcoin would bring it into the eyes of the mainstream media, with more attention comes more miners using already purchased hardware to mine. In attempting to take over the network within any reasonable span of time a nation would bring enough attention to itself to increase the security of the network. The fact is they just dont make the high-end cards fast enough for them to get their hands on even 30,000 within a year.

To run a mining farm of  30.000 superpumped 3x6990 rigs (that will do 3X the actual hash power) you need at least 500 full time employees (in 6 month at least 30 million $), (Godaddy has 1500 technical employees for 60.000 servers)
There are energy bill to pay even if is a governative installation (at least 20Million $ for 6 month paying 0,05$/KWh). Add a good 10 million $ of network infrastructure and connectivity (I don't calculate a redundant 10GBit connection to internet assuming a govern can obtain it almost for free) . UPS and backup generator will cost you another 50-60 million $ (we are talkin of a 90MW, a small power plant - a similar one, highly ecological,  built 4-5 years ago costs about 200M$). The cooler must be really efficient to work in a similar environment (one rig produce 1,5KW of heat, like a powerful server with 8CPU and 24 disks. But no server farm hosts 30.000 of that servers for what I know ): at least 30M$.   
The 30.000 rigs wil cost at least 40 million $ (I've considered a 20%-25% less than the wholesaler prices, probably lower than the price you can have buying a bigger stock directly from producers, and with no assembly cost).
Total: 200M$, plus the building, the connectivity, the security (both from hackers and physical assaults). If you're in hurry to build it you can easily double the costs.
Overall if a govern wants  to kill bitcoin, it's still a small investment to do.

So, no one can get the cards that fast, and even if they could facility costs alone would be 10 times more than your estimate. Now assuming that the attacking group was patient they would wait for a few years while they build up the needed supply of cards, of course buying that many cards would bring attention to Bitcoin, and with that attention comes more mining, meaning that their goal becomes higher the more cards they try to buy, and there is no guarantee that the network would not move to dedicated hardware that trumps graphics cards when they where in the middle of collection.

The only way to escape this paradox and get the cards within any reasonable time spawn would be to build their own factory dedicated to the production of high end gpus or dedicated hardware.  Building a factory is complicated, they require clean rooms and millions in state of the art equipment, as well as large amounts of expensive materials to build. Producing their own cards may be the only way to preform this attack without alerting us, but it is also much much more expensive.


Both approaches require lots of prep time, one to build up cards the other to build factories.

[cunicula]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

No single company has the resources to kill off Bitcoin, a government with a few hundred million could build a card manufacturing plant, and pump out the cards in the needed volume. But that would take years of preparation. Even if they started today no entity on earth could take over the block chain in less than 6 months.

This is ridiculous. Ebay owns paypal and has a market cap of 43 billion. That is about 4000 times as much money as the appromiximate 10 million required to disrupt block processing. If they want to they could stop bitcoin within a month. It is chump change to them. If they took bitcoin as a serious threat they would probably have been killed it off already.

It would be a lot more than that

Cost is not the limiting factor when it comes to preforming a 51% attack, taking over the block chain is cheap, the problem is getting the cards. As it is the small mining community has already made a major impact in the supply of high end ATI cards, it would take months for any person to get their hands on enough cards to preform a serious attack. Any attempt to acquire such a large number of cards so quickly would throw up red flags all over the place and give the nation under attack ample time to prepare. Any nation attacking Bitcoin would bring it into the eyes of the mainstream media, with more attention comes more miners using already purchased hardware to mine. In attempting to take over the network within any reasonable span of time a nation would bring enough attention to itself to increase the security of the network. The fact is they just dont make the high-end cards fast enough for them to get their hands on even 30,000 within a year.

To run a mining farm of  30.000 superpumped 3x6990 rigs (that will do 3X the actual hash power) you need at least 500 full time employees (in 6 month at least 30 million $), (Godaddy has 1500 technical employees for 60.000 servers)
There are energy bill to pay even if is a governative installation (at least 20Million $ for 6 month paying 0,05$/KWh). Add a good 10 million $ of network infrastructure and connectivity (I don't calculate a redundant 10GBit connection to internet assuming a govern can obtain it almost for free) . UPS and backup generator will cost you another 50-60 million $ (we are talkin of a 90MW, a small power plant - a similar one, highly ecological,  built 4-5 years ago costs about 200M$). The cooler must be really efficient to work in a similar environment (one rig produce 1,5KW of heat, like a powerful server with 8CPU and 24 disks. But no server farm hosts 30.000 of that servers for what I know ): at least 30M$.    
The 30.000 rigs wil cost at least 40 million $ (I've considered a 20%-25% less than the wholesaler prices, probably lower than the price you can have buying a bigger stock directly from producers, and with no assembly cost).
Total: 200M$, plus the building, the connectivity, the security (both from hackers and physical assaults). If you're in hurry to build it you can easily double the costs.
Overall if a govern wants  to kill bitcoin, it's still a small investment to do.

So, no one can get the cards that fast, and even if they could facility costs alone would be 10 times more than your estimate. Now assuming that the attacking group was patient they would wait for a few years while they build up the needed supply of cards, of course buying that many cards would bring attention to Bitcoin, and with that attention comes more mining, meaning that their goal becomes higher the more cards they try to buy, and there is no guarantee that the network would not move to dedicated hardware that trumps graphics cards when they where in the middle of collection.

The only way to escape this paradox and get the cards within any reasonable time spawn would be to build their own factory dedicated to the production of high end gpus or dedicated hardware.  Building a factory is complicated, they require clean rooms and millions in state of the art equipment, as well as large amounts of expensive materials to build. Producing their own cards may be the only way to preform this attack without alerting us, but it is also much much more expensive.


Both approaches require lots of prep time, one to build up cards the other to build factories.

ttk2, I am not going to respond to you anymore.

[asdf]

Couldn't any institution buy all the cards on the first and second hand market? This would drive up the price enticing more sellers (many probably honest miners). They would quickly buy 30,000 cards, albeit at greatly inflated prices?

Think about it. If a 6990 is going for $2000, how many people are just going to cash in their mining rig?

[ttk2]

Is intentional sabotage illegal in any wa?One scenario is that a company sabotages bitcoin in order to kill off competition. Paypal might consider this for example. What is the weakest link in terms of sabotaging bitcoin? In particular is it easier to target the.mining process or verificatoon by the bitcoin client?

No single company has the resources to kill off Bitcoin, a government with a few hundred million could build a card manufacturing plant, and pump out the cards in the needed volume. But that would take years of preparation. Even if they started today no entity on earth could take over the block chain in less than 6 months.

This is ridiculous. Ebay owns paypal and has a market cap of 43 billion. That is about 4000 times as much money as the appromiximate 10 million required to disrupt block processing. If they want to they could stop bitcoin within a month. It is chump change to them. If they took bitcoin as a serious threat they would probably have been killed it off already.

It would be a lot more than that

Cost is not the limiting factor when it comes to preforming a 51% attack, taking over the block chain is cheap, the problem is getting the cards. As it is the small mining community has already made a major impact in the supply of high end ATI cards, it would take months for any person to get their hands on enough cards to preform a serious attack. Any attempt to acquire such a large number of cards so quickly would throw up red flags all over the place and give the nation under attack ample time to prepare. Any nation attacking Bitcoin would bring it into the eyes of the mainstream media, with more attention comes more miners using already purchased hardware to mine. In attempting to take over the network within any reasonable span of time a nation would bring enough attention to itself to increase the security of the network. The fact is they just dont make the high-end cards fast enough for them to get their hands on even 30,000 within a year.

To run a mining farm of  30.000 superpumped 3x6990 rigs (that will do 3X the actual hash power) you need at least 500 full time employees (in 6 month at least 30 million $), (Godaddy has 1500 technical employees for 60.000 servers)
There are energy bill to pay even if is a governative installation (at least 20Million $ for 6 month paying 0,05$/KWh). Add a good 10 million $ of network infrastructure and connectivity (I don't calculate a redundant 10GBit connection to internet assuming a govern can obtain it almost for free) . UPS and backup generator will cost you another 50-60 million $ (we are talkin of a 90MW, a small power plant - a similar one, highly ecological,  built 4-5 years ago costs about 200M$). The cooler must be really efficient to work in a similar environment (one rig produce 1,5KW of heat, like a powerful server with 8CPU and 24 disks. But no server farm hosts 30.000 of that servers for what I know ): at least 30M$.    
The 30.000 rigs wil cost at least 40 million $ (I've considered a 20%-25% less than the wholesaler prices, probably lower than the price you can have buying a bigger stock directly from producers, and with no assembly cost).
Total: 200M$, plus the building, the connectivity, the security (both from hackers and physical assaults). If you're in hurry to build it you can easily double the costs.
Overall if a govern wants  to kill bitcoin, it's still a small investment to do.

So, no one can get the cards that fast, and even if they could facility costs alone would be 10 times more than your estimate. Now assuming that the attacking group was patient they would wait for a few years while they build up the needed supply of cards, of course buying that many cards would bring attention to Bitcoin, and with that attention comes more mining, meaning that their goal becomes higher the more cards they try to buy, and there is no guarantee that the network would not move to dedicated hardware that trumps graphics cards when they where in the middle of collection.

The only way to escape this paradox and get the cards within any reasonable time spawn would be to build their own factory dedicated to the production of high end gpus or dedicated hardware.  Building a factory is complicated, they require clean rooms and millions in state of the art equipment, as well as large amounts of expensive materials to build. Producing their own cards may be the only way to preform this attack without alerting us, but it is also much much more expensive.


Both approaches require lots of prep time, one to build up cards the other to build factories.

ttk2, I am not going to respond to you anymore.

Not sure if i should feel victorious or insulted.

[jtimon]

I have created this thread to find out if bitcoin could survive only with fees.

[makomk]

So, no one can get the cards that fast, and even if they could facility costs alone would be 10 times more than your estimate. Now assuming that the attacking group was patient they would wait for a few years while they build up the needed supply of cards, of course buying that many cards would bring attention to Bitcoin, and with that attention comes more mining, meaning that their goal becomes higher the more cards they try to buy, and there is no guarantee that the network would not move to dedicated hardware that trumps graphics cards when they where in the middle of collection.

The only way to escape this paradox and get the cards within any reasonable time spawn would be to build their own factory dedicated to the production of high end gpus or dedicated hardware.  Building a factory is complicated, they require clean rooms and millions in state of the art equipment, as well as large amounts of expensive materials to build. Producing their own cards may be the only way to preform this attack without alerting us, but it is also much much more expensive.

Alternatively they could just manufacture some structured ASICs and rely on commercial confidentiality provisions in their contracts to hide exactly what they're doing until it's too late. (If we're considering the possibility of stealthy US government attackers, it's even worse: they have deals with foundries in the US for ASIC manufacturing to handle classified projects.)

[ttk2]

So, no one can get the cards that fast, and even if they could facility costs alone would be 10 times more than your estimate. Now assuming that the attacking group was patient they would wait for a few years while they build up the needed supply of cards, of course buying that many cards would bring attention to Bitcoin, and with that attention comes more mining, meaning that their goal becomes higher the more cards they try to buy, and there is no guarantee that the network would not move to dedicated hardware that trumps graphics cards when they where in the middle of collection.

The only way to escape this paradox and get the cards within any reasonable time spawn would be to build their own factory dedicated to the production of high end gpus or dedicated hardware.  Building a factory is complicated, they require clean rooms and millions in state of the art equipment, as well as large amounts of expensive materials to build. Producing their own cards may be the only way to preform this attack without alerting us, but it is also much much more expensive.

Alternatively they could just manufacture some structured ASICs and rely on commercial confidentiality provisions in their contracts to hide exactly what they're doing until it's too late. (If we're considering the possibility of stealthy US government attackers, it's even worse: they have deals with foundries in the US for ASIC manufacturing to handle classified projects.)

I never said an attack was impossible, yes this method would work, but like a said, it would still take a long time to make enough devices, and it would still be very very very expensive.

[Boussac]

I think we'll learn more about this aspect of BitCoin at the start of next year, when inflation will halve.

I guess "inflation" here means "money supply growth" because bitcoin offers a deflationary model.
Incidentally, my undestanding was that the period was four years so BTC generation will halve only at the end of next year, right ??

This thread deals with two issues, namely robustness against possible attacks AND sustainability of mining when new bitcoin generation dwindles to zero. Even if the two issues are connected I would like to focus on the second one, the economics of mining.

If K₀ is the generation rate of the bitcoins in the initial period (300 Bitcoins/hour) and P the duration of each period (4 years), according to the rules of the bitcoin protocol Qn = P*K₀/2ⁿ is the quantity of bitcoins generated in the n^th period.
In other words, the aggregate value of the money created in the nth period equals the aggregate value of the money created in the first four years (the initial period) if  the value of the bitcoins grows by an annual rate of 2ⁿ/4*n.

Now let us assume the utility of the bitcoin network is proportional to the square of N the number of users (a formula to factor in the “network effect”).
Assuming the utility of the bitcoin network correlates to the value of the bitcoins yields:
(N_n)² = k * 2^n-2/n.

The result shows that an exponential growth of the user base, i.e. a continued successful launch of the bitcoin network in line with what we are currently witnessing, ensures a steady growth of the value generated by mining at least until an inflexion point is reached.

[Meni Rosenfeld]

Incidentally, my undestanding was that the period was four years so BTC generation will halve only at the end of next year, right ??

It's roughly 4 years, but when difficulty is increasing it will be less. I think the current estimate is September 2012.

Now let us assume the utility of the bitcoin network is proportional to the square of N the number of users (a formula to factor in the “network effect”).

There's some network effect, but the utility still scales much slower than quadratic.

Assuming the utility of the bitcoin network correlates to the value of the bitcoins yields:
(N_n)² = k * 2^n-2/n.

The result shows that an exponential growth of the user base, i.e. a continued successful launch of the bitcoin network in line with what we are currently witnessing, ensures a steady growth of the value generated by mining at least until an inflexion point is reached.

We know there's no problem as long as generation is high enough... But what happens afterwards?

[Boussac]

My point was there is no problem as long as user base grows exponentially.
Afterwards (that is after the inflexion point), commission fees must be the partial substitute for value and could maintain the profitability of mining in the long run by virtue of deflation.

[DiamondPlus]

It doesn't make sense as long as people still follow the rule of "build off of whichever version of a block you see first", but if some people stop doing that then some possibilities open up.

Imaging the slow version was offering the finder of the bock who built on top of it 1BTC. Why not switch over once you get the slow block? This could move the slow blocks chance of staying from 5% to 50% if half of miners were looking for this and it would be well worth it to pay that 1BTC.

Once something like this is in place you have the option of working on the current block which has very few fees in it or redoing the previous block to get all those fees and paying people to work on yours instead. If the offering you'll need to pay (just enough to outbid the other finder) is less than the extra fees then this is what you do.

I don't think people will keep working on old blocks and waste a ton of work, I think an equilibrium fee of about half the average fees in a block will emerge. At half there is no incentive to work on an old block because you'll have to pay more than half to outbid the other guy who is making the standard offering.

I think this is cool because your tx fee would now indirectly go to pay for more and more work getting piled on instead of just the one block of work as it is now.

-DiamondPlus

[jtimon]

It doesn't make sense as long as people still follow the rule of "build off of whichever version of a block you see first", but if some people stop doing that then some possibilities open up.

I think this is not true.
I think the rule is: "hash on top of the longest chain (measured in work, not in blocks)".

[Gavin Andresen]

It doesn't make sense as long as people still follow the rule of "build off of whichever version of a block you see first", but if some people stop doing that then some possibilities open up.

I think this is not true.
I think the rule is: "hash on top of the longest chain (measured in work, not in blocks)".

Yes, but if two blocks are found with the same difficulty then they are considered equal work-wise (even if one has a smaller hash value-- the "Bits" field in the block is used to compute chain work, not the block hash).

If there are several valid tip-of-block-chain choices all of which have the same work, then miners are free to build off whichever one they like. The rule everybody (I think!) is using now is "build off the one seen first."

[jtimon]

I think this is not true.
I think the rule is: "hash on top of the longest chain (measured in work, not in blocks)".

Yes, but if two blocks are found with the same difficulty then they are considered equal work-wise (even if one has a smaller hash value-- the "Bits" field in the block is used to compute chain work, not the block hash).

Thanks for the clarification.

If there are several valid tip-of-block-chain choices all of which have the same work, then miners are free to build off whichever one they like. The rule everybody (I think!) is using now is "build off the one seen first."

I see. The "first seen" rule cannot be enforced in the protocol because peers cannot know what the next block solver received first. But there's no incentive to use other rule.

[ripper234]

A recent post from Mike Hearn about how to handle this problem with network assurance contracts.

[DeathAndTaxes]

Hmm, I fail to see why the fees would drop with time.  The upper limit on the rate of transaction processing should drive the fees up.

From this post (of mine) http://bitcointalk.org/index.php?topic=5758.msg84788#msg84788

Just some rough math to answer my own question, a 1MB block would contain roughly 3000 typical transactions.  Thus there's an upper limit on the transaction processing rate for the entire economy - I reckon that means that once the transaction rate rises over an average of 18k/hr., there will be a minimum fee for your transaction to even have a chance of being processed, and market forces will drive the minimum fee up with time.

The demand for performing transactions will presumably continue to increase, but the supply is constant, ~18k/hr.

Block sizes will need to get larger in time.  3000 transactions per block and one block every 600 seconds = max limit of 5tps.  To put it into perspective Paypal handles about 50 tps.  However transactions are "bursty" so daily avg peak is likely 100 tps to 200 tps.  The annual peak is probably in the 400 to 500 tps range. 

Also the 3000 transactions is based on avg current size.  As transactions get larger and more complex 1MB will result in less transactions.  At 2400 transactions you are looking at ~4tps, 1800 transactions is a mere 3tps.

Bitcoin supports larger block sizes and they will be used if Bitcoin ever scales to any significant volume.

Fees are a problem IMHO (I know Gavin disagrees).  It creates a race to the bottom. 

PROCESSING A BLOCK IS VERY DIFFICULT = high cost per block.
PROCESSING A TRANSACITON IS VERY EASY = low cost per transaction.

This creates a dynamic where it makes no economical sense for a miner to exclude paying transactions as long as there is room in the block.  Keeping the block size artificially small would drive up prices BUT it also limits the usefulness of the network (max tps) and there will be a never ending push to increase block sizes. 

So the efficient miner includes all paying transactions and excludes all free ones.  That maximizes the revenue per unit of work. Remember 99.99999999999999999999999% of the work/cost is hashing the block excluding a paying transaction drops your revenue by magnitudes more than it drops your work/cost.

So if miners include all paying transactions and block sizes will need to grow to allow the network to grow the optimal fee to pay is the lowest fee which gets you in the next block.  That will trend towards 1 satoshi per transaction.  Fees will rise as the transaction volume pushes up against the limit of block size but then fall again as block size is increased. 

At even paypal scale (50 tps) and 1 satoshi per transaction annual revenue for network is $80.00.