Soft block size limit reached, action required by YOU

[Daily Anarchist]

While I highly favor increasing the block size I also like the idea of full nodes being compensated somehow. Miners get rewarded, those who don't run full nodes get rewarded, but full nodes currently bear the cost out of almost pure altruism. That can't last.

[SgtSpike]

The fee system was created to prevent transaction spam.... use it!

[niko]

Repeat... Let's keep the soft block size limit at 250KB and see what will happen  

I don't think people are so stupid that without that added transaction capacity they will leave bitcoin

Yes.  Bitcoin is an experiment, and a complex one for that matter. This is a rare opportunity to see how this complex system responds to limited block size, limited bandwidth of an average node, and relatively large volume of transactions.

This thread is one of responses, and even though I dislike the idea, I appreciate the opportunity to observe how it unfolds. We should all watch and learn, and only act after carefully thinking through all the intended and unintended consequences of our actions.

At this point, only an idiot would have a firm opinion about the interplay of SD, dust transactions, economics of mining and fees, scaling, Moore's Law, and future adoption rate, modes of use, and value of coins.

[bullioner]

lol @ Ian's comment.

Anyway, I'll just point out that mining generates three things that are beneficial:

1) Mining revenue (inflation+fees)
2) Heat
3) Network security

Assume revenue drops to zero. This won't happen until we're all dead, but OK, let's think about it anyway.

It is still better to use mining hardware for heat rather than just resistance coils, because confirmations are a useful thing. Somebody, somewhere, will find a way to pay for that. Some random ideas:

 - Exchanges giving out free mining hardware to elderly people
 - Companies that rely on irreversible transactions heating their offices with mining hardware
 - Insurance companies that secure transactions on the behalf of other companies selling the waste heat to swimming pool complexes or shopping malls

In case you're being serious, this is unrealistic.  For one thing, mining will continue to become relatively more capital intensive; rather than the costs of mining being primarily from electricity, they will be from the capital investment in equipment required.

Another thing: Bitcoin mining is not the first industry with the potential for portable units and which  produces large amounts of waste heat.  To estimate how much heat might be recoverable economically, and what the value might be relative to input energy costs, one could survey existing industries.  Anyway this approach is going to be seasonal.

It's basic capitalism - it's an activity that benefits all kinds of people, so it'll get funded one way or another.  Now whether that is by attaching fees to every tx (as Satoshi felt might be one day needed), or network assurance contracts, or simply finding ways to profit off the waste heat ... this is an interesting academic debate but as I said, inflation only drops to zero after our bones have all turned to dust. There's plenty of time to sit back and see what happens.

The time when it drops to zero is not significant.  The lower the rewards are, the easier attacks get, continuously.  Pointing out that the very tail end of the exponential decay is well in the future is a red herring.  We know that the current rate of 10% or so annually is probably more than the long-term rate required.  In around 7 years, after the next two rewards halvings, the reward rate will be more like 2% of the monetary base annually, without transaction fees.  At that point, I would be seriously worried if that were the only direct incentive to miners.  

[Sunny King]

Everybody isn't cut from the same mold.  Different folks have different limits.

I was running a full node on a vendor's VPS service (roughtly equivalent to an Amazon Micro Instance).  [Very good connectivity, no bandwidth limit, good data rate.]  As of a week or two ago, it can no longer handle the volume of transactions.  Virtual memory is being exhausted.  I'm not willing to pay the higher fees for the next level of service.  That node is now gone.

Can user set a limit on memory pool size? Then you wouldn't run into this problem I assume? Did you manage to see how much ram it was using at the time?

[Jutarul]

Bottleneck is not hardware, but bandwidth. Average joe's internet connection has to be able to handle a full node. ( 1Mbit/s w/ 10Mbit/s peak)

Everybody isn't cut from the same mold.  Different folks have different limits.

I was running a full node on a vendor's VPS service (roughtly equivalent to an Amazon Micro Instance).  [Very good connectivity, no bandwidth limit, good data rate.]  ...

Why do I have the feeling that it's a bad idea to host your validation node in the cloud?

[Syke]

The fee system was created to prevent transaction spam.... use it!

Bring back the .01 minimum fee. Problem solved.

[zebedee]

hazek I don't understand you here.  I normally agree with the vast majority of what you say.  You're very pro free-market and normally rational.  

But on this one issue you seem to think it needs to be solved by wise men rather than market prices.  Why the lack of faith on this one issue?  No I don't believe there is a tragedy of the commons here.

Because I don't want the security of Bitcoin to fall victim to the tragedy of the commons or get centralized between a few super nodes. And if that is the path the market will take Bitcoin on, I will unfortunately need to move to on to better things.

It's not obvious though that having no limit, and encouraging more participation and paying transactions, wouldn't give rise to more fees than limiting it and making the system less functionally attractive.

What your own philosophy should find offensive about your own suggestion though is that you're trying to push the cost for something you want (higher perceived security) onto others by artificially limiting the number of txns in a block.

If people like you as a group want a higher hashrate, they will be prepared to pay for it, and pay the market rate, whatever it is.

Perhaps people sitting with 1,000 BTC at an address should feel a slight need to transfer it back to themselves occasionally, paying a fraction of a BTC txn fee, to increase the security of their BTC rather than "coast" on the fees of others paying for transactions?  Or get a network assurance contract like Mike suggests.  Personally I suspect the network will be more than able to sustain itself on fees, even when the block reward runs out, provided there's enough space for the txns that people want to do.

These problems will be resolved by those who care about it paying for it.

[niko]

The fee system was created to prevent transaction spam.... use it!

Bring back the .01 minimum fee. Problem solved.

Fixing the minimum fee in bitcoins for everyone is a bad idea, as you don't know how value of coins will change, but you do know it's been changing wildly. This is something that is best left to the free market - miners and payers - to figure out and adjust as we go.

[hazek]

hazek I don't understand you here.  I normally agree with the vast majority of what you say.  You're very pro free-market and normally rational.  

But on this one issue you seem to think it needs to be solved by wise men rather than market prices.  Why the lack of faith on this one issue?  No I don't believe there is a tragedy of the commons here.

Because I don't want the security of Bitcoin to fall victim to the tragedy of the commons or get centralized between a few super nodes. And if that is the path the market will take Bitcoin on, I will unfortunately need to move to on to better things.

It's not obvious though that having no limit, and encouraging more participation and paying transactions, wouldn't give rise to more fees than limiting it and making the system less functionally attractive.

What your own philosophy should find offensive about your own suggestion though is that you're trying to push the cost for something you want (higher perceived security) onto others by artificially limiting the number of txns in a block.

If people like you as a group want a higher hashrate, they will be prepared to pay for it, and pay the market rate, whatever it is.

Perhaps people sitting with 1,000 BTC at an address should feel a slight need to transfer it back to themselves occasionally, paying a fraction of a BTC txn fee, to increase the security of their BTC rather than "coast" on the fees of others paying for transactions?  Or get a network assurance contract like Mike suggests.  Personally I suspect the network will be more than able to sustain itself on fees, even when the block reward runs out, provided there's enough space for the txns that people want to do.

These problems will be resolved by those who care about it paying for it.

You are confused. I don't want anything but what I and everyone else already agreed to. The client I downloaded and consented to use hears only those who obey the limit of 1Mb, I simply don't want that to change right now because I think strong arguments have been made that it would lead to bad outcomes. Everyone else already agreed to these rules and it's them who want to have me to agree to a change of these rules, not the other way around.

So my attitude towards this problem is perfectly inline with the principles I hold dearly and above all else.


But if it was me who was redesigning Bitcoin and I had the options to set fresh rules that anyone could choose to agree to, you're damn right I'd set the rules that would incentivize users to pay for security. In the long run it's the only way security will be provided for, if users pay for it.

Also security affects those who make a fresh transaction the most so it makes sense they are the ones who pay for it.

[ProfMac]

Seriously Mike? People will use miners for heat? Man for a smart engineer you sure say some pretty stupid things.

We have a wind turbine and -20°F temperatures.
Why ship electricity when we can have heat and ship bitcoin?

[Mike Hearn]

In case you're being serious, this is unrealistic.  For one thing, mining will continue to become relatively more capital intensive; rather than the costs of mining being primarily from electricity, they will be from the capital investment in equipment required.

I am being serious, although I am not claiming that this actually will happen. Just that it's one of many possibilities, and we don't really know how funding of network security will play out over the long run.

Your estimates about capital equipment expenditure don't seem right to me. Mining hardware just isn't that expensive, especially once ASIC companies start getting economies of scale. We're very close to the limit of what mining technology can do before it's only a few process nodes behind state of the art.

Currently, cost of mining is going up because it's transitioning from an activity for hobbyists to an activity for professionals who need specialized equipment. There is an equilibrium point at which the cost of slightly faster hardware is not balanced by the increased revenue.

Another thing: Bitcoin mining is not the first industry with the potential for portable units and which  produces large amounts of waste heat.  To estimate how much heat might be recoverable economically, and what the value might be relative to input energy costs, one could survey existing industries.  Anyway this approach is going to be seasonal.

Yes it will be somewhat seasonal. Mining is ideally suited to this though because you don't really need anything beyond a source of electricity (wind, waves, biogas, whatever), some hardware and an internet connection. In particular it means you can extract energy from places where it's uneconomical to get it all the way to the grid.

Also, it's worth remembering that all large electricity grids need load balancing capacity. Today, grid operators have contracts with large consumers of power such that they can remotely shut down or spin up heavy equipment to balance moment-to-moment fluctuations in demand. For example, in the United Kingdom there's a hydropower facility called Dinorwig which pumps water up a mountain or lets it flow down in order to absorb/generate power spikes. Storage heaters and the Economy 7 tariff also incentivise people to use electricity when there's a temporary surplus and avoid it when there's a temporary shortfall.

Mining, being an activity you can switch on or off instantly, is an ideal form of load balancing - if there's a sudden unexpected drop in electricity demand for some reason, or just an anticipated drop at night, mining hardware can switch on and absorb the excess capacity reducing the need for power shedding or inefficient cooldowns of gas turbines.

The time when it drops to zero is not significant.  The lower the rewards are, the easier attacks get, continuously. Pointing out that the very tail end of the exponential decay is well in the future is a red herring.  We know that the current rate of 10% or so annually is probably more than the long-term rate required.  In around 7 years, after the next two rewards halvings, the reward rate will be more like 2% of the monetary base annually, without transaction fees.  At that point, I would be seriously worried if that were the only direct incentive to miners.  

You don't know what will happen to the exchange rate in future. Currently the FX rate is quite clearly the dominating factor in how much mining is done. An inflation of 0.5 BTC per block could be more than enough if each coin is worth thousands of dollars.

[bullioner]

You don't know what will happen to the exchange rate in future. Currently the FX rate is quite clearly the dominating factor in how much mining is done. An inflation of 0.5 BTC per block could be more than enough if each coin is worth thousands of dollars.

True that they could each have that value, but try looking at it this way instead.  Incentives for a systemic attack on Bitcoin rise with the total value of the monetary base.  Therefore, it makes sense to reason about incentives for securing the transaction log in terms of a rate expressed in terms of the total Bitcoin monetary base, or the total Bitcoin transaction volume, or some combination.  If those incentives just seem high expressed in dollars, but are a dwindling fraction of the total value of the Bitcoin system, we enter a stage of increased risk to the system the lower those incentives go as a fraction of the system's value.  Leaving aside for the moment what I'll slightly unfairly call externalities (assurance bonds et al), do you have some sympathy with the view that incentives to secure the system should grow with the value of the system, and should therefore be considered in terms expressed using the system's parameters?

I find the metric of the percentage of the Bitcoin monetary base spent annually on securing the transaction log nice, because it's easy for people to reason about.  Talking about 0.5 BTC per block with each BTC worth thousands of dollars doesn't, I think, make it easy for people to reason about whether they think the equilibrium achieved with such incentives would be sufficient to secure the network.  Expressing the same thing in a different way: that around 0.15% of the system's value would go annually on this direct incentive for securing the system, I think, gives a statistic that's more meaningful.  And this meaning scales with the value of the system as a whole.

I don't think anyone claims to know what that risk distribution is, but intuitively I would guess a few percent per annum of the total monetary base should go annually on direct incentives to secure the transaction log.  And as far as I'm concerned, this means transaction fees taking over from minting as the source of this direct incentive.

For the record, I'm not against the block size growing, perhaps hugely.  I just think that ensuring adequate transaction fees has to remain part of the design.  One way of doing this is to tie increases in max block size to market pressure, over previous blocks, for that to happen.  12 BTC per block will equate to a few percent of Bitcoin's monetary base per year, by the way.

[Mike Hearn]

Incentives for a systemic attack on Bitcoin rise with the total value of the monetary base.

I don't think that's correct. You can't attack bitcoin as a whole, that's not what majority attacks let you do. You can only attempt to defraud individual merchants, either by double spending or by finding a dumb merchant that isn't using a full node.

The actual maximum incentive to attack Bitcoin is the value of the most expensive thing you can steal by doing a double spend minus the cost of doing it. Therefore the right amount of mining is controlled by the observed double spend rate against active merchants not the size of the monetary base.


 

[muyuu]

Incentives for a systemic attack on Bitcoin rise with the total value of the monetary base.

I don't think that's correct. You can't attack bitcoin as a whole, that's not what majority attacks let you do. You can only attempt to defraud individual merchants, either by double spending or by finding a dumb merchant that isn't using a full node.

The actual maximum incentive to attack Bitcoin is the value of the most expensive thing you can steal by doing a double spend minus the cost of doing it. Therefore the right amount of mining is controlled by the observed double spend rate against active merchants not the size of the monetary base.

That's right, in practice the biggest attacks would be the size of a "big" single expenditure. And however big it is, it's going to be tiny in comparison to the monetary mass, and in reality it's not going to grow in "real value terms" over time.

The incentive for miners is not related to the monetary mass, but to transactional volume.

[zebedee]

Incentives for a systemic attack on Bitcoin rise with the total value of the monetary base.

I don't think that's correct. You can't attack bitcoin as a whole, that's not what majority attacks let you do. You can only attempt to defraud individual merchants, either by double spending or by finding a dumb merchant that isn't using a full node.
 

I agree with everything else you've said Mike apart from this.  Majority attacks let you kill confidence in the network (either by not letting transactions get processed, or ripping off a few merchants) which if not resolved would ultimately bring down the network itself.  No?

[muyuu]

Incentives for a systemic attack on Bitcoin rise with the total value of the monetary base.

I don't think that's correct. You can't attack bitcoin as a whole, that's not what majority attacks let you do. You can only attempt to defraud individual merchants, either by double spending or by finding a dumb merchant that isn't using a full node.
 

I agree with everything else you've said Mike apart from this.  Majority attacks let you kill confidence in the network (either by not letting transactions get processed, or ripping off a few merchants) which if not resolved would ultimately bring down the network itself.  No?

Possibly, but that's not related nor proportional to the monetary base.

Some people seem to have the idea that there are attacks capable of executing transactions from arbitrary addresses. That's not the case, you still need a valid transaction and that's only possible with the source address' private key. Double spending attacks are limited to whatever quantity the attacker had in the first place, it doesn't matter how many coins are there in total.

[Notanon]

Yes it will be somewhat seasonal. Mining is ideally suited to this though because you don't really need anything beyond a source of electricity (wind, waves, biogas, whatever), some hardware and an internet connection. In particular it means you can extract energy from places where it's uneconomical to get it all the way to the grid.

Also, it's worth remembering that all large electricity grids need load balancing capacity. Today, grid operators have contracts with large consumers of power such that they can remotely shut down or spin up heavy equipment to balance moment-to-moment fluctuations in demand. For example, in the United Kingdom there's a hydropower facility called Dinorwig which pumps water up a mountain or lets it flow down in order to absorb/generate power spikes. Storage heaters and the Economy 7 tariff also incentivise people to use electricity when there's a temporary surplus and avoid it when there's a temporary shortfall.

Mining, being an activity you can switch on or off instantly, is an ideal form of load balancing - if there's a sudden unexpected drop in electricity demand for some reason, or just an anticipated drop at night, mining hardware can switch on and absorb the excess capacity reducing the need for power shedding or inefficient cooldowns of gas turbines.

Another thing to consider is if a miner is either on a flat tariff for their electricity use or if they are on a time-of-use tariff. With the latter, running a mining rig overnight during the offpeak period might be the most economical way of mining, but I'm not aware of any mining software that can tell the rig to switch on at a given time and power down accordingly just before the offpeak period finishes.

Another possibility is with a house equipped with PV panels, the user could add an option to interface with the energy meter or home automation monitor that would run the miner while the PV generation capacity is producing an average net amount minus standby power consumption and maybe an airconditioning load (air conditioner size and PV system output pending) and run it then. Use with wind turbines might be a bit trickier, though, given the erratic nature of wind, though I believe in conjunction with a battery storage system it may be more viable and could allow the mining rig to run off the battery storage system alone.

[Mike Hearn]

I agree with everything else you've said Mike apart from this.  Majority attacks let you kill confidence in the network (either by not letting transactions get processed, or ripping off a few merchants) which if not resolved would ultimately bring down the network itself.  No?

In practice you wouldn't be building a parallel chain all the time, just when you want to execute an attack. And I'm pretty much expecting double spends to happen from time to time in future, maybe even a lot. For some classes of goods (think virtual goods or goods where theft is easily resolved by the police), it'll make more sense to just absorb a bit of fraud than pay for additional mining.

[muyuu]

I agree with everything else you've said Mike apart from this.  Majority attacks let you kill confidence in the network (either by not letting transactions get processed, or ripping off a few merchants) which if not resolved would ultimately bring down the network itself.  No?

In practice you wouldn't be building a parallel chain all the time, just when you want to execute an attack. And I'm pretty much expecting double spends to happen from time to time in future, maybe even a lot. For some classes of goods (think virtual goods or goods where theft is easily resolved by the police), it'll make more sense to just absorb a bit of fraud than pay for additional mining.

If in the future we have double-spend attacks happening "from time to time" I think Bitcoin would have failed to find a balance good enough to stay secure. Not enough base difficulty to make double-spend attacks unfeasible??

If trust is necessary then that's the end of it.