Arguments in the Bitcoin Block Size Debate(davidsterry.com)
davidsterry.com
Arguments in the Bitcoin Block Size Debate
http://davidsterry.com/blog/2015/06/arguments-in-the-bitcoin-block-size-debate/
5 comments
"Unless > 50% of the miners, by hash capacity, make a change that affects the blockchain, it doesn't happen."
This is not true at all. If the top 4 pools start accepting and mining 20MB blocks, no one else in the network will accept the blocks, even if that chain is longer, because it would violate the fundamental rule, hard coded in their Bitcoin implementation, that blocks must be < 1MB. It would effectively fork the block chain.
This is not true at all. If the top 4 pools start accepting and mining 20MB blocks, no one else in the network will accept the blocks, even if that chain is longer, because it would violate the fundamental rule, hard coded in their Bitcoin implementation, that blocks must be < 1MB. It would effectively fork the block chain.
> So the free market will solve any overload problem.
What do you mean by this? If 2mb of transaction data is generated every 10 minutes, but only 1mb of it is mined, what is the free market solution?
If the goal is to have bitcoin be an actual working currency for the people (and it may not be that to you), then it would seem the goal should be to be able to process the vast majority of transactions, instead of just those for the richest half.
A "solution" that is economic and solved by free markets tends to be about how to distrubute limited resources. I.e. here's a nice piece of property for sale - who gets it? The highest bidder. problem solved.
But I don't think the challenge for bitcoin is how to auction off space in the blockchain, it's how to make that space wide enough to be reliable as a potential currency for common people. If, every time you bought a candy bar at walgreens, you had to pay for the candy bar and take a guess at an extra fee for your payment and hope it goes through 10 minutes later... it just would not work.
So the market would be solving something, but it wouldn't be towards the goal of bitcoin being a currency. And honestly I think that's what the opponents of the block size increase want. As cryptocurrency tech evolves, we're seeing that various parties have new self interests to promote. Competing technologies, etc. And they benefit by bitcoin becoming artificially limited. I know there are other reasons, but it's a big part of why this is even a debate.
What do you mean by this? If 2mb of transaction data is generated every 10 minutes, but only 1mb of it is mined, what is the free market solution?
If the goal is to have bitcoin be an actual working currency for the people (and it may not be that to you), then it would seem the goal should be to be able to process the vast majority of transactions, instead of just those for the richest half.
A "solution" that is economic and solved by free markets tends to be about how to distrubute limited resources. I.e. here's a nice piece of property for sale - who gets it? The highest bidder. problem solved.
But I don't think the challenge for bitcoin is how to auction off space in the blockchain, it's how to make that space wide enough to be reliable as a potential currency for common people. If, every time you bought a candy bar at walgreens, you had to pay for the candy bar and take a guess at an extra fee for your payment and hope it goes through 10 minutes later... it just would not work.
So the market would be solving something, but it wouldn't be towards the goal of bitcoin being a currency. And honestly I think that's what the opponents of the block size increase want. As cryptocurrency tech evolves, we're seeing that various parties have new self interests to promote. Competing technologies, etc. And they benefit by bitcoin becoming artificially limited. I know there are other reasons, but it's a big part of why this is even a debate.
"What do you mean by this? If 2mb of transaction data is generated every 10 minutes, but only 1mb of it is mined, what is the free market solution?"
When you submit a transaction, you're entering an auction. The winners of the auction get their transactions processed quickly. The losers don't. The market thus sets a price on getting a transaction processed. This discourages free-riders who expect their transactions to be processed free or for a nominal fee. Miners are facing the dreaded "halving" next year, when the block reward drops by half. They need higher fees to compensate.[1]
[1] http://www.coindesk.com/new-study-low-bitcoin-transaction-fe...
When you submit a transaction, you're entering an auction. The winners of the auction get their transactions processed quickly. The losers don't. The market thus sets a price on getting a transaction processed. This discourages free-riders who expect their transactions to be processed free or for a nominal fee. Miners are facing the dreaded "halving" next year, when the block reward drops by half. They need higher fees to compensate.[1]
[1] http://www.coindesk.com/new-study-low-bitcoin-transaction-fe...
This isn't an actual solution though, because space is still limited. Bitcoin cannot be a currency if it can only fit 2,000 transactions per block. Even if everyone is willing to pay $100 in fees for their $1 transactions, they won't all fit, and that's broken.
The challenge is to find a way for miners to be incentivized and for them also to process all reasonable transactions. There's no reason miner's couldn't increase their fees independently of the block size increase.
The challenge is to find a way for miners to be incentivized and for them also to process all reasonable transactions. There's no reason miner's couldn't increase their fees independently of the block size increase.
> They need higher fees to compensate.
That's not an argument for any block size in particular, though. Saying that fee revenue is maximized at smaller blocks is positing that demand for block space (i.e. transaction demand) is relatively inelastic, whereas there isn't really any consensus or even much discussion on what the elasticity is.
That's not an argument for any block size in particular, though. Saying that fee revenue is maximized at smaller blocks is positing that demand for block space (i.e. transaction demand) is relatively inelastic, whereas there isn't really any consensus or even much discussion on what the elasticity is.
Auctions for limited capacity at this ridiculously early stage in bitcoin's growth will shake confidence in the network to the point of abandonment.
A currency you can't easily transact in is one nobody wants to use. An auction process like this means you can't easily use bitcoin.
A currency you can't easily transact in is one nobody wants to use. An auction process like this means you can't easily use bitcoin.
the goal should be to be able to process the vast majority of transactions
That implies the social value of the transactions outweighs the cost of carrying them on the blockchain. If the bulk of transactions are penny-valued "scratch card" gambles (are they still a thing?), that is obviously not true.
That implies the social value of the transactions outweighs the cost of carrying them on the blockchain. If the bulk of transactions are penny-valued "scratch card" gambles (are they still a thing?), that is obviously not true.
I don't think the miners have as much say as you think. If the users all get together and hard fork the protocol, either the miners go along with it, or their blocks will be rejected, no matter how much hashing power they have.
The only thing they could reasonably do is go rogue and try to 51% attack the new fork. If they did that successfully, however, they'd be showing that bitcoin isn't as strong as people think it is, thereby destroying their own business (it really is practically being mined by only 5 to 10 miners nowadays).
This is why I like alternative coins like Peercoin and Neucoin which are trying to iron out the details of proof-of-stake. Even Ethereum is supposedly going to move to a proof-of-stake model.
The only thing they could reasonably do is go rogue and try to 51% attack the new fork. If they did that successfully, however, they'd be showing that bitcoin isn't as strong as people think it is, thereby destroying their own business (it really is practically being mined by only 5 to 10 miners nowadays).
This is why I like alternative coins like Peercoin and Neucoin which are trying to iron out the details of proof-of-stake. Even Ethereum is supposedly going to move to a proof-of-stake model.
The miners can refuse to make 20MB blocks. Unless the hard fork makes 1MB blocks invalid, which isn't being suggested.
With less than 50% of miners, you can't force much of a backwards-compatible hard fork like this one.
I had a discussion about this recently here https://www.reddit.com/r/Bitcoin/comments/38xfw3/forking_don...
With less than 50% of miners, you can't force much of a backwards-compatible hard fork like this one.
I had a discussion about this recently here https://www.reddit.com/r/Bitcoin/comments/38xfw3/forking_don...
The miners are the ones creating blocks. Users (i.e. transaction senders) simply have to accept it. They have no input into making or verifying blocks.
This is not true. People who run full nodes (as a group) have just as much power as miners. If they all run a hard-forked client, then the miners will be happily mining away on their (worthless) chain.
That just isn't true. The miners are the one that decide which chain is "real" and which isn't.
The frontend has code to verify the decision of the miner, and it has no ability to change anything.
The frontend has code to verify the decision of the miner, and it has no ability to change anything.
>The frontend has code to verify the decision of the miner, and it has no ability to change anything.
But, if all the users agree to use frontend code that rejects the decision of the miners that refuse to play along, then the chain the miners are working on is effectively dead.
But, if all the users agree to use frontend code that rejects the decision of the miners that refuse to play along, then the chain the miners are working on is effectively dead.
>The miners are users, so all users won't agree to anything adverse to that set of miners.
The miners aren't trading any goods or services for their bitcoin. If merchants and other major players in the bitcoin ecomony go along with the hard fork, the previous chain is dead.
The miners aren't trading any goods or services for their bitcoin. If merchants and other major players in the bitcoin ecomony go along with the hard fork, the previous chain is dead.
The miners are users, so all users won't agree to anything adverse to that set of miners.
How are the miners going to get users to switch over to their frontend? How are miners going to get companies like bitpay and coinbase to switch to their fork? It's up to the users to decide the value of the coins in each fork.
It's irrelevant as to how many miners are working on the original chain. Mining could be done on a single laptop by one person for all the protocol cares.
Lets say that 99.9% percentage of the miners decided not to go along with a hard fork, and 0.1% did, effectively bitcoin would be split up into 2 networks, A and B.
Functionally, they would be nearly equivalent, ignoring network lag. As a user you could even have two clients, one that accepts transactions for network A, and one for network B. And your wallet in network A could even have a different balance than in network B. So how do you decide which wallet is correct?
The same way everyone else does, by mass consensus, ie. by what everyone else is using. Ex, if you wanted to buy stuff from merchant X, and they had a network B wallet, it wouldn't matter how many miners were mining on network A, would it?
Miners only provide security against a double spend 51% attack and that's it. By their very nature, they also are in the best position to do a 51% attack. Therefore the only leverage they have is they could threaten the new network. But as I was explaining before, if they attack network B, it goes to show that network A is nearly as vulnerable as network B is.
In other words, its as centralized as fiat, which destroys one of the most highly touted features of Bitcoin as a currency of the people. If a couple of a bad actors (ie the people running the mining warehouses), can destroy it, either because of government coercion or by their own free will, then as you can see, Bitcoin is like the emperor who has no clothes. It could be destroyed at any time.
So, we have a MAD situation (Mutually Assured Destruction). If the users force the miners hand too much, the miners could retaliate, by destroying bitcoin by executing a 51% attack. But of course, as long as the miners are making money, they don't want to do that.
Nobody invested in the bitcoin wants anyone to know this, the users nor the miners, because it goes against their best interest to show the weakness of the currency they own.
However, the longer this draws on, the more people will wake up and see that Bitcoin is truly becoming centralized into something that resembled Beenz and Floodz from the 90's. Not a currency operated by the people, but a currency owned and operated by a few large corporations.
Lets say that 99.9% percentage of the miners decided not to go along with a hard fork, and 0.1% did, effectively bitcoin would be split up into 2 networks, A and B.
Functionally, they would be nearly equivalent, ignoring network lag. As a user you could even have two clients, one that accepts transactions for network A, and one for network B. And your wallet in network A could even have a different balance than in network B. So how do you decide which wallet is correct?
The same way everyone else does, by mass consensus, ie. by what everyone else is using. Ex, if you wanted to buy stuff from merchant X, and they had a network B wallet, it wouldn't matter how many miners were mining on network A, would it?
Miners only provide security against a double spend 51% attack and that's it. By their very nature, they also are in the best position to do a 51% attack. Therefore the only leverage they have is they could threaten the new network. But as I was explaining before, if they attack network B, it goes to show that network A is nearly as vulnerable as network B is.
In other words, its as centralized as fiat, which destroys one of the most highly touted features of Bitcoin as a currency of the people. If a couple of a bad actors (ie the people running the mining warehouses), can destroy it, either because of government coercion or by their own free will, then as you can see, Bitcoin is like the emperor who has no clothes. It could be destroyed at any time.
So, we have a MAD situation (Mutually Assured Destruction). If the users force the miners hand too much, the miners could retaliate, by destroying bitcoin by executing a 51% attack. But of course, as long as the miners are making money, they don't want to do that.
Nobody invested in the bitcoin wants anyone to know this, the users nor the miners, because it goes against their best interest to show the weakness of the currency they own.
However, the longer this draws on, the more people will wake up and see that Bitcoin is truly becoming centralized into something that resembled Beenz and Floodz from the 90's. Not a currency operated by the people, but a currency owned and operated by a few large corporations.
Miners with 51% could make only 1MB blocks and orphan all larger ones.
Your reasoning is only valid in cases where blocks from one fork aren't accepted on the other, but this isn't true for blocksize.
Your reasoning is only valid in cases where blocks from one fork aren't accepted on the other, but this isn't true for blocksize.
This is off-topic, but bitlaunder.com is arguably a poor choice for Bitcoin mixing, given that it does business in the US. Far better is Bitcoin Fog at foggeddriztrcar2.onion, because coercing them will be harder.
What do you mean by "does business in the US"?
Per their website:
Per their website:
HEAD OFFICE LOCATION:
Moscow Citydel
9 Zemlyanoyval, 4th Floor
Moscow 105064
Russian Federation
+74959679797Sorry, my mistake. Perhaps I was just spreading gossip.
But do see http://bitcoin.stackexchange.com/questions/25722/bitlaunders...
But do see http://bitcoin.stackexchange.com/questions/25722/bitlaunders...
In my view, it's a great and simple short term fix, waiting for more volumes and bigger bucks to get involved. I like it.
I have a fairly reductionist view on this, one I think is helpful, so here goes.
First, I think of almost all large-scale Bitcoin arguments as religious. Especially if you're an outsider to Bitcoin, it will probably help you to think of them that way, too.
People come to a belief, and then back-fill plausible arguments all the time. When there are vigorous debates like this in Bitcoin land, we usually have some strongly held underlying beliefs coming to conflict.
I would say the holy war here revolves around a single axiom on each side: the "Small Block" camp fundamentally believes that there is a sort of inflation that happens when block sizes grow. Just like they believed that increased adoption would increase Bitcoin prices, they believe at some level that increased scarcity for block slots will increase transaction fees, and that the system relies on this promise to maintain its integrity.
The "Large Block" camp fundamentally believes that the value in the protocol comes from cheap, large-scale transaction support, and that artificially limiting right now is harmful.
You can reduce most of the arguments to comments on this. In particular, I think the comments about centralized mining concerns are almost totally specious. Bitcoin has long ago left behind small hobbyist miners with tiny data connections. Block size changes would not have helped them keep up -- there are fundamental aggregation economics in mining that have pushed them out. Any small miner that wants to mine now will do exactly the same thing they would do with 1 Gigabyte blocks: get their work from a pool, and spot verify they haven't been cheated.
Embedded in the world view of most "Large Block" miners if you follow out the logic is the belief that probably transaction fees won't pay for the entire security of the network. Ed Felten put out some calculations to this effect while he was at Princeton. They're reasonably compelling arguments, or at least worth thinking hard about. If you give up on the idea that tx fees have to pay for the entire security of the network, I think large blocks become a super obvious choice to enhance the current utility of the network.
If you do not want to give up on that idea, then any block size increase decreases the chance that the tx fee scheme will work. Since it is currently not working (the last two blocks as I write this had .09 and .02 BTC in transaction fees respectively), all such changes must be vigorously fought.
At any rate, this is why the "small block" folks have a 'purity' vibe when they argue; I think they feel they are protecting the original scheme, one they bought into and care about.
As you might be able to tell, I don't think the tx fee scheme is working right now and hence am a "large block" kind of guy. Forcing it to work, (e.g. shrinking block sizes dramatically, and perhaps refusing to mine on blocks with insufficient fees) would dramatically lower the usefulness of Bitcoin to many people. As it is, it's super annoying to have to wonder if a transaction will get included in the next block.
Finally, there are some possible tx fee solutions orthogonal to block size miners could implement easily. I have thought for a number of years that miners should just have 'express, regular, bulk' time/fee cutoffs for txs they would include. They should publish these rates. Of course, this could only be done by larger pools, or a consortium of smaller ones. Even 20% would make an impact on the fee schedule.
When I think about the economics though, I ask myself 'who would it help?' Miners are already dealing with very difficult economics at a macro and micro level, especially if they have chip projects. Getting a little more juice out of their mining would be nice, but is absolutely not what the large-scale mining game is about. On the flip side, it adds cognitive load and frustration to people transacting. It doesn't make a ton of sense to me.
If we're in a world where Bitcoin is still struggling to be useful to a billion people daily, then cognitive load and frustration should be ruthlessly excised from the technology until we get to critical mass.
First, I think of almost all large-scale Bitcoin arguments as religious. Especially if you're an outsider to Bitcoin, it will probably help you to think of them that way, too.
People come to a belief, and then back-fill plausible arguments all the time. When there are vigorous debates like this in Bitcoin land, we usually have some strongly held underlying beliefs coming to conflict.
I would say the holy war here revolves around a single axiom on each side: the "Small Block" camp fundamentally believes that there is a sort of inflation that happens when block sizes grow. Just like they believed that increased adoption would increase Bitcoin prices, they believe at some level that increased scarcity for block slots will increase transaction fees, and that the system relies on this promise to maintain its integrity.
The "Large Block" camp fundamentally believes that the value in the protocol comes from cheap, large-scale transaction support, and that artificially limiting right now is harmful.
You can reduce most of the arguments to comments on this. In particular, I think the comments about centralized mining concerns are almost totally specious. Bitcoin has long ago left behind small hobbyist miners with tiny data connections. Block size changes would not have helped them keep up -- there are fundamental aggregation economics in mining that have pushed them out. Any small miner that wants to mine now will do exactly the same thing they would do with 1 Gigabyte blocks: get their work from a pool, and spot verify they haven't been cheated.
Embedded in the world view of most "Large Block" miners if you follow out the logic is the belief that probably transaction fees won't pay for the entire security of the network. Ed Felten put out some calculations to this effect while he was at Princeton. They're reasonably compelling arguments, or at least worth thinking hard about. If you give up on the idea that tx fees have to pay for the entire security of the network, I think large blocks become a super obvious choice to enhance the current utility of the network.
If you do not want to give up on that idea, then any block size increase decreases the chance that the tx fee scheme will work. Since it is currently not working (the last two blocks as I write this had .09 and .02 BTC in transaction fees respectively), all such changes must be vigorously fought.
At any rate, this is why the "small block" folks have a 'purity' vibe when they argue; I think they feel they are protecting the original scheme, one they bought into and care about.
As you might be able to tell, I don't think the tx fee scheme is working right now and hence am a "large block" kind of guy. Forcing it to work, (e.g. shrinking block sizes dramatically, and perhaps refusing to mine on blocks with insufficient fees) would dramatically lower the usefulness of Bitcoin to many people. As it is, it's super annoying to have to wonder if a transaction will get included in the next block.
Finally, there are some possible tx fee solutions orthogonal to block size miners could implement easily. I have thought for a number of years that miners should just have 'express, regular, bulk' time/fee cutoffs for txs they would include. They should publish these rates. Of course, this could only be done by larger pools, or a consortium of smaller ones. Even 20% would make an impact on the fee schedule.
When I think about the economics though, I ask myself 'who would it help?' Miners are already dealing with very difficult economics at a macro and micro level, especially if they have chip projects. Getting a little more juice out of their mining would be nice, but is absolutely not what the large-scale mining game is about. On the flip side, it adds cognitive load and frustration to people transacting. It doesn't make a ton of sense to me.
If we're in a world where Bitcoin is still struggling to be useful to a billion people daily, then cognitive load and frustration should be ruthlessly excised from the technology until we get to critical mass.
There's an argument to be made that pools doesn't mean centralized, because miners can switch to a different pool if a pool is abusive. Whether that actually works is another debate.
That has happened in the past, when the P2SH proposal was batted around, pools voted, and people did switch based on their politics.
More often though, miners will have a main pool and backup pools based on uptime.
In the scrypt mining world, pool hopping happened (happens?) for economic reasons; hash -> litecoin variant -> USD (or BTC)-denominated return is continually optimized. This caused some wild swings in mining difficulty for some scrypt-based coins as their exchange rates changed.
More often though, miners will have a main pool and backup pools based on uptime.
In the scrypt mining world, pool hopping happened (happens?) for economic reasons; hash -> litecoin variant -> USD (or BTC)-denominated return is continually optimized. This caused some wild swings in mining difficulty for some scrypt-based coins as their exchange rates changed.
> the last two blocks as I write this had .09 and .02 BTC in transaction fees respectively
What's that in real money?
Less trollishly, how does it compare to the mining reward? To the mining reward 10 years from now?
What's that in real money?
Less trollishly, how does it compare to the mining reward? To the mining reward 10 years from now?
>Less trollishly, how does it compare to the mining reward?
Mining reward is 25BTC, so on average 1/500.
>To the mining reward 10 years from now?
https://en.bitcoin.it/wiki/Controlled_supply
3.125 BTC.
Mining reward is 25BTC, so on average 1/500.
>To the mining reward 10 years from now?
https://en.bitcoin.it/wiki/Controlled_supply
3.125 BTC.
About $25....combined.
Does the author have any evidence for his claim "The limit was temporary"? It was certainly commonly claimed even a couple of years ago that the coming scarcity of blockchain space would lead to an increase in fees.
https://bitcointalk.org/index.php?topic=287.msg8810#msg8810
>The eventual solution will be to not care how big it gets.
https://bitcointalk.org/index.php?topic=1347.msg15366#msg153...
https://bitcointalk.org/index.php?topic=946236.msg10388435#m...
>I'm the guy who went over the blockchain stuff in Satoshi's first cut of the bitcoin code. Satoshi didn't have a 1MB limit in it. The limit was originally Hal Finney's idea. Both Satoshi and I objected that it wouldn't scale at 1MB. Hal was concerned about a potential DoS attack though, and after discussion, Satoshi agreed. The 1MB limit was there by the time Bitcoin launched. But all 3 of us agreed that 1MB had to be temporary because it would never scale.
>The eventual solution will be to not care how big it gets.
https://bitcointalk.org/index.php?topic=1347.msg15366#msg153...
https://bitcointalk.org/index.php?topic=946236.msg10388435#m...
>I'm the guy who went over the blockchain stuff in Satoshi's first cut of the bitcoin code. Satoshi didn't have a 1MB limit in it. The limit was originally Hal Finney's idea. Both Satoshi and I objected that it wouldn't scale at 1MB. Hal was concerned about a potential DoS attack though, and after discussion, Satoshi agreed. The 1MB limit was there by the time Bitcoin launched. But all 3 of us agreed that 1MB had to be temporary because it would never scale.
Ah, that's really significant, thanks
Relevant: coverage of various perspectives and positions by Ethan Heilman along with counterarguments to each:
https://github.com/EthanHeilman/BlockSizeDebate/blob/master/...
I like that it also includes discussion of the Lightning Network's role, and the idea of making the block size dynamic. These are less heard of perspectives that nevertheless seem significant.
https://github.com/EthanHeilman/BlockSizeDebate/blob/master/...
I like that it also includes discussion of the Lightning Network's role, and the idea of making the block size dynamic. These are less heard of perspectives that nevertheless seem significant.
If the blockchain runs out of capacity, the transactions with the biggest fee attached get processed first. So the free market will solve any overload problem.
Bitcoin volume in dollars isn't increasing.[2] The peak was over a year ago. The number of transactions is up, but not the total dollar value. Some of the transactions are just "mixing" transactions, to assist in money laundering. The companies that offer such services are quite open about this.[3]
[1] https://blockchain.info/pools [2] https://blockchain.info/charts/estimated-transaction-volume-... [3] https://bitlaunder.com/