What are the downsides to 8MB blocks?

[maokoto]

If a bigger block affects microtransactions, that is really bad. I think that the ability of sending a fraction of a cent is one of the good things of Bitcoin (microearnings, crowdfunding by lots of people, etc). Is it really that so? Bigger blocks will hinder microtransactions?

[uxgpf]

If a bigger block affects microtransactions, that is really bad. I think that the ability of sending a fraction of a cent is one of the good things of Bitcoin (microearnings, crowdfunding by lots of people, etc). Is it really that so? Bigger blocks will hinder microtransactions?

No. It's the opposite.

When transaction volume bumps into blocksize limit, it causes increase in fees, which can make microtransactions unfeasible. (You might have to pay more in fees than your transaction is worth) Fees rise as people compete to get their transactions included to the block.

With bigger blocksize limit there's no or very little fee competition and thus microtransactions remain feasible.

[poeEDgar]

If a bigger block affects microtransactions, that is really bad. I think that the ability of sending a fraction of a cent is one of the good things of Bitcoin (microearnings, crowdfunding by lots of people, etc). Is it really that so? Bigger blocks will hinder microtransactions?

No. It's the opposite.

When transaction volume bumps into blocksize limit, it causes increase in fees, which can make microtransactions unfeasible. (You might have to pay more in fees than your transaction is worth) Fees rise as people have to compete to get their transactions included to the block.

With bigger blocksize limit there's no or very little fee competition and thus microtransactions remain feasible.

Do you think we should address dust transactions which maximize output/transaction size (i.e. spam)? After all, competition for inclusion in blocks during stress tests is absolutely not a result of organic growth in transaction volume (adoption). This is obvious, since it diverges sharply from average transaction growth, then reverts to the mean following the stress test.

Perhaps increasing the block size is not the only way to address spam attacks.....

[Linuld]

Linuld, really... you are asking the same questions that have been answered like a thousand times, and making arguments that have been countered a thousand times...

I will only comment on some of them:

On top... you know that the current blocks are not even filled full. They only are full when these spammers act. And there is no reason to assume that suddenly, with 8 Megabyte Blocks, these blocks will be full. Where should all these transactions come from?

There's a perfect reason to assume that blocks are suddenly 8Mb -- if there's a specific attack vector linked to full 8Mb blocks, it will be used. It's not that expensive, especially when fees are gravitating towards infignificant amounts due to big blocks.

8 Megabytes per 10 Minutes and you think we will get in problems with that? What kind of internet connection do you have in order to fear that?

A decentralized network robustness depends not on an average throughput, but on the throughput of bottlenecks, namely network bottlenecks (China-ROW, e.g.), CPU bottlenecks and others.
CPU bottlenecks are currently being worked on by core devs.

Sure. And will you pay that? You realize the cost of the recent spam attacks? And those were only some KB. You speak about filling up a lot MB with spam. I wonder if you will find someone who will do the spamming.

Do you realize that yeaterday's attack created more than 35Mb transaction backlog -- enough to fill more than four 8Mb blocks.
The previous attack had much larger backlog.

And of course i referred to the natural adoption of bitcoin that will fill the blocks. Spamming won't happen since it does not make sense then anymore in any way for the spammer.

"Sense"? What's the sense for an attacker currently? Do you realize that with this natural adoption even 8Mb blocks will get full at some point, and the closer they get to 8Mb, the less costly is a spam attack.
We can't solve the spam attack by allowing for more bloating, we can fight it only with fees, which are bound to go towards zero if we keep increasing blocksize limit beyond demand.

What do you write here? You assume they will be filled because you assume it will be so? You know, i know that the past has shown that you are not correct. So i'm fine with me being not convinced by your assumption.

The bottlenecks you describe are nonsense too. The bitcoin network can't remain so lowtech that each and every modem user can use it. It only must be useable with the hardwar that is widely used. And that includes relatively old hardware too. But when someone uses tech from 10 years ago then he can't await that he can use every software running nowadays. I mean you set higher rules then they are usual in ever other area. And that makes no sense.

*lol* You argument with 35 MB backlog? And? Then will be four 8 MB blocks filled. Then the spook is over.

And again... do you want to spam a network with 8 MB blocks? For what reason?

*lol* Again an argument that you use that will be solved with 8 MB blocks. You say if the blocks are nearly full attacks will cost more and that we should not upgrade to 8 MB because they eventually will be full too at some point. So let's keep 1MB? Surely that makes sense.

If we fight the spam with fees then bitcoin will lose one of it's rare advantages. But i guess you would be ok for that? Altcoin involved or so?

But maybe you don't want the blocksize rise because it is only a temporary fix. Then yes, i would prefer no limit at all. Past has shown that it was no problem to have enough free space in the blocks. They did not fill up because some ominous spammers with deep pockets filled them up.

[Linuld]

If we increase this to 2 txs, and the second tx doesn't appear (spare space), then the first tx will simply lower its fees (blocksize becomes non-scarce).

That's again an assumption that was proven wrong long ago by reality. As long as the 1 MB blocks were not full the fees did not vanish like you make the story go. The bitcoiners still pay their fee. Might be because of the wrong assumption that it will go much faster or is needed. In fact it is a little bit faster and the feest are very small anyway. So no one cares really paying them.

Your assumption is already proven incorrect. There is no reason to assume that will change only because we have 50% free space in a 8 MB block instead in a 1MB block.

BUT

After our increase, we doubled the amount of users that can make transactions in 10min intervals.

We doubled the amount of possible transactions, but we can't magically double the actual amount of them.

What's the point of Bitcoin if you can only send money to a limited amount of people?

That's not my point. I want Bitcoin to scale while still retaining its unique characteristics.

Yes, we doubled the amount of possible transactions. And that's something crucially needed. Imagine paypal having a max global allowed transaction amount per hour? That would be plain stupid and everyone would go away.

If suddenly a big company like ebay would start accepting bitcoin then bitcoin would show it is useless. No one would use it after having seen how limited it is. Artificially limited.

It's good that you want to keep bitcoins unique characteristics but surely higher fees are not the way. On top it would mean unreliablity since when you somehow end with a fee too low, you will not get it send. That is deadly for a payment system.

So sorry, your way will surely kill bitcoin in the long run.

[Linuld]

A good argument against the bitcoiners that demand 1MB blocksize because "Miners have to be rewarded" is the electricity that goes away with that. I recently read an article about a study saying that currently 1! bitcoin transaction costs so much electricity like 1.75 normal US households use in a day.

We can't let all these miners survive artificially by raising the fees. The amount of miners need to be trimmed down and the bitcoin security be more power efficient. That was always happening and it should and can't be held up artificially. The bitcoin network would be still very very secure with less miners.

Unfortunately miners are the mighty ones in the game. We only can hope that the sane one are the biggest ones. And that they see the need of bitcoin adoption for getting future profits.

That is a sick amount of power... So mining one block consumes the power of a full city? That is insane and not sustainable, especially if it keeps rising...

I do agree that's a lot to secure this level of Bitcoin transactions as of today. But I also read that current amount of hashing power is enough to secure even a 100 fold increase in the transaction number and user number. So to say, if we would have 100 million users, the network would be secured with today's hash rate.

That's why I don't understand the opponents of the block size increase. Bitcoin has to scale in order to be sustainable in the long run! For that we need bigger blocks!

but WHO will ever be able to download 8MB

8 FUCKING MASSIVE MEGA BYTES

this will surely make minning WAY more centralized

Well then we have a huge freaking problem! Then Bitcoin isn't ready for mass adoption and how will ever be ready? What solution is there besides sidechains, if the sidechains are even possible?

I have a typical home connection downloading 8MB might take me anywhere from 2 to 6 seconds

if i'm streaming a movie it could take 12seconds i guess

so it clearly unacceptable to expect miners to download 8MB, so there can be no doubt 8MB WILL centralize bitcoin mining.

You think we have decentralized mining now? Where did you life the last year? Oo

And you fear 2-6 seconds download time? What is so much better in needing 0.25 to 0.75 seconds? What will it safe for you?

And it will take a long time with the normal adoption rate of bitcoin to reach 8MB constantly.

What exactly is your problem with that? It's not like there will something negative happen when you don't get a randomly spread block in around 10 minutes in under 1 second.

[Linuld]

A good argument against the bitcoiners that demand 1MB blocksize because "Miners have to be rewarded" is the electricity that goes away with that. I recently read an article about a study saying that currently 1! bitcoin transaction costs so much electricity like 1.75 normal US households use in a day.

We can't let all these miners survive artificially by raising the fees. The amount of miners need to be trimmed down and the bitcoin security be more power efficient. That was always happening and it should and can't be held up artificially. The bitcoin network would be still very very secure with less miners.

Unfortunately miners are the mighty ones in the game. We only can hope that the sane one are the biggest ones. And that they see the need of bitcoin adoption for getting future profits.

That is a sick amount of power... So mining one block consumes the power of a full city? That is insane and not sustainable, especially if it keeps rising...

I do agree that's a lot to secure this level of Bitcoin transactions as of today. But I also read that current amount of hashing power is enough to secure even a 100 fold increase in the transaction number and user number. So to say, if we would have 100 million users, the network would be secured with today's hash rate.

That's why I don't understand the opponents of the block size increase. Bitcoin has to scale in order to be sustainable in the long run! For that we need bigger blocks!

but WHO will ever be able to download 8MB

8 FUCKING MASSIVE MEGA BYTES

this will surely make minning WAY more centralized

Well then we have a huge freaking problem! Then Bitcoin isn't ready for mass adoption and how will ever be ready? What solution is there besides sidechains, if the sidechains are even possible?

I have a typical home connection downloading 8MB might take me anywhere from 2 to 6 seconds

if i'm streaming a movie it could take 12seconds i guess

so it clearly unacceptable to expect miners to download 8MB, so there can be no doubt 8MB WILL centralize bitcoin mining.

Are you being serious here Adam or ....?

I also have a typical home connection and 8MB would take me about 0.5-1 sec, if I would be a miner, I wouldn't stream a movie on the connection, but buy a second one.
If my current connection is to slow I still have the possibility to upgrade to a higher speed connection.

So if I can do it, surely a miner making millions should be able to figure it out.

if miners can't stream 3 HD pron movies all at once while mining on a typical home connections , mining will become centralized! simple as that.

Sure, because there is no way to specify the mining client to be prioritized. I mean bitcoin must be so low tech that every porn streamer can mine and stream at the same time.

That is so very much arbitrarely, it's funny somehow. I mean why not lets say that we can't take out the commodore 64 users? Let's downgrade bitcoin because we don't want anyone being discriminated. It's not the fault of the person who does hinder bitcoin on it's work, right?

I await the very few remaining home miners are a lot smarter than that.

Again... problems come up that are no problems.

[Linuld]

No worries. Hearn says that miners whose bandwidth is too limited can create their own altcoin and leave bitcoin.

No worries. The remaining core developers say that bitcoiners that think that bitcoin fees became way to costly, transctions going through are uncertain and so on can use the great new altcoin sidechain lightning network and leave bitcoin.



Oh the irony.

How i hate how the bitcoin development has become a pool of politics that don't work on the main thing anymore but on their own little side project that they work for in reality. Corruption i would name that in the real world.

It's a pity.

[Linuld]

If a bigger block affects microtransactions, that is really bad. I think that the ability of sending a fraction of a cent is one of the good things of Bitcoin (microearnings, crowdfunding by lots of people, etc). Is it really that so? Bigger blocks will hinder microtransactions?

No. It's the opposite.

When transaction volume bumps into blocksize limit, it causes increase in fees, which can make microtransactions unfeasible. (You might have to pay more in fees than your transaction is worth) Fees rise as people have to compete to get their transactions included to the block.

With bigger blocksize limit there's no or very little fee competition and thus microtransactions remain feasible.

Do you think we should address dust transactions which maximize output/transaction size (i.e. spam)? After all, competition for inclusion in blocks during stress tests is absolutely not a result of organic growth in transaction volume (adoption). This is obvious, since it diverges sharply from average transaction growth, then reverts to the mean following the stress test.

Perhaps increasing the block size is not the only way to address spam attacks.....

Though the result of an artificial block size limit would nothing less than raising fees when the blocks are full most of the time. So microtransactions won't work anymore the way they were. So it doesn't really matter if the spam attacks artifically raised fees. With 1MB blocks blocks will be full all the time and one point. And that inevitably means rising fees.

As if the bitcoin network needs rising fees. Mining is so rewarding at the moment that we have 100 times more hashingpower than needed to secure the network. 100 times. Which translates to 1 transaction costing so much power like 1.75 average us households use in one whole day.

Sure, i can see why miners want to earn more or want their old miners still earn. But it is completely impossible to fight the blockhalving with rising fees too.

[RoadTrain]

What do you write here? You assume they will be filled because you assume it will be so? You know, i know that the past has shown that you are not correct. So i'm fine with me being not convinced by your assumption.

No, I assume they will be filled in order to model system's behaviour under these circumstances -- a part of engineering work, that you're downplaying under the premise that blocks aren't gonna be 8MB.
Let me maker it clear: blocks won't be full just because -- is a weak and pointless assumption. Good engineering must take care of extremes -- if we can support 8MB, then let's make sure the systems stays robust.

The bottlenecks you describe are nonsense too. The bitcoin network can't remain so lowtech that each and every modem user can use it. It only must be useable with the hardwar that is widely used. And that includes relatively old hardware too. But when someone uses tech from 10 years ago then he can't await that he can use every software running nowadays. I mean you set higher rules then they are usual in ever other area. And that makes no sense.

Don't exaggerate and twist my argument. Bottlenecks matter, especially network bottlenecks between miners, as they affect block propagation.

*lol* You argument with 35 MB backlog? And? Then will be four 8 MB blocks filled. Then the spook is over.

And again... do you want to spam a network with 8 MB blocks? For what reason?

Once again, if the system is deemed unstable under 8MB blocks, then there's a clear reason for an attack.

*lol* Again an argument that you use that will be solved with 8 MB blocks. You say if the blocks are nearly full attacks will cost more and that we should not upgrade to 8 MB because they eventually will be full too at some point. So let's keep 1MB? Surely that makes sense.

Solved? Care to read my post once more? It's not solved, it's kicking the can down the road.

If we fight the spam with fees then bitcoin will lose one of it's rare advantages. But i guess you would be ok for that? Altcoin involved or so?

What's this advantage? Low fees? I guess not, it's main unique advantage is trustlessness. It holds a lot more value than cheap transactions, which can be routinely provided by many centralized solutions. And if it's under threat because of your idea that Bitcoin must be effectively free, I'd rather stick with trustlessness.
I wonder how you are gonna fund miners when the block subsidy diminishes. Haven't thought about that?

But maybe you don't want the blocksize rise because it is only a temporary fix. Then yes, i would prefer no limit at all. Past has shown that it was no problem to have enough free space in the blocks. They did not fill up because some ominous spammers with deep pockets filled them up.

Jeez. The only thing past has shown us is that miners don't care about transaction fees much because of subsidy, while users don't bother adjusting their default fees. This will change in the future as fees become more significant relative to the block subsidy. With no limit, fees will gravitate to arbitrarily low near-zero values. I wonder how miners are gonna be funded. By 8GB blocks? No thanks, I don't think I'm interested in this kind of centralized Bitcoin.

[RoadTrain]

If we increase this to 2 txs, and the second tx doesn't appear (spare space), then the first tx will simply lower its fees (blocksize becomes non-scarce).

That's again an assumption that was proven wrong long ago by reality. As long as the 1 MB blocks were not full the fees did not vanish like you make the story go. The bitcoiners still pay their fee. Might be because of the wrong assumption that it will go much faster or is needed. In fact it is a little bit faster and the feest are very small anyway. So no one cares really paying them.

Your assumption is already proven incorrect. There is no reason to assume that will change only because we have 50% free space in a 8 MB block instead in a 1MB block.

False. The only reason why the first tx would still pay 1BTC fee is that because the owner didn't bother adjusting his fee policy. That's what's happening now with Bitcoin. The sole reason for it is block subsidy. As it diminishes, fees become more significant.

BUT

After our increase, we doubled the amount of users that can make transactions in 10min intervals.

We doubled the amount of possible transactions, but we can't magically double the actual amount of them.

What's the point of Bitcoin if you can only send money to a limited amount of people?

That's not my point. I want Bitcoin to scale while still retaining its unique characteristics.

Yes, we doubled the amount of possible transactions. And that's something crucially needed. Imagine paypal having a max global allowed transaction amount per hour? That would be plain stupid and everyone would go away.

Bitcoin is nothing like centralized Paypal and can't be scaled the same way. Pointless comparison.

It's good that you want to keep bitcoins unique characteristics but surely higher fees are not the way. On top it would mean unreliablity since when you somehow end with a fee too low, you will not get it send. That is deadly for a payment system.

Do you know that it's exactly how Bitcoin is functioning? If you pay a fee too low, your tx will likely end up stuck.

So sorry, your way will surely kill bitcoin in the long run.

That's a very bold statement, especially when you don't know 'my way'.

[coalitionfor8mb]

Good engineering must take care of extremes -- if we can support 8MB, then let's make sure the systems stays robust.

In the foreseeable future any occasional big blocks will likely get orphaned, so your concern is unfounded.
However, by the time the blocks are consistently big, the technology will already have improved to be able to handle them.
With static limit, big blocks now will become small blocks in the future and the process will have to repeat itself.

This is a good way to keep the system in check, allow it grow, while at the same time not let it spiral out of our hands.

[adamstgBit]

if miners can't stream 3 HD pron movies all at once while mining on a typical home connections , mining will become centralized! simple as that.

Sure, because there is no way to specify the mining client to be prioritized. I mean bitcoin must be so low tech that every porn streamer can mine and stream at the same time.

yes! bitcoin must be so low tech that every porn streamer can mine and stream at the same time.

or else i consider bitcoin mining centralized.

also if i can't get gramma into minning same result  i consider bitcoin mining centralized.

i believe to be considered decentralized, mining must be accessible to EVERYONE and i mean everyone. point final!


if you haven't caught on yet i'm just trolling the 1MBter's argument the requiring miners to download >1MB blocks will somehow centralize mining.

[adamstgBit]

No worries. Hearn says that miners whose bandwidth is too limited can create their own altcoin and leave bitcoin.

why don't they just join a pool? better yet they can keep mining that wtv pool they are currently mining at.

[brg444]

No worries. Hearn says that miners whose bandwidth is too limited can create their own altcoin and leave bitcoin.

why don't they just join a pool? better yet they can keep mining that wtv pool they are currently mining at.

do you realize what this entails?

these guys have more than 50% of the hashing power!

of course that is what they are going to do: centralize.

centralize to improve their connectivity and abuse larger blocks by pushing the biggest ones they can make down the throat of other miners.

[johnyj]

No worries. Hearn says that miners whose bandwidth is too limited can create their own altcoin and leave bitcoin.

No worries, miners whose bandwidth is almost unlimited can create their own altcoin and leave bitcoin  

A faster bandwidth coped with large blocks will separate you from majority of nodes, because majority of the nodes will not receive your block before they mined the next block (each relay of the block between two nodes will add 3-30 seconds delay depends on the network bandwidth, and there are many hops for a block to reach majority of nodes), so your block is always orphaned by the rest of the network, means your only choice is to fork your own coin

In fact, the size of the block is not limited by either core devs or miners, but ISPs around the world. I believe without major upgrade of internet infrastructure, the bitcoin network can not handle 2MB blocks well, not even mention 8MB

[RoadTrain]

Good engineering must take care of extremes -- if we can support 8MB, then let's make sure the systems stays robust.

In the foreseeable future any occasional big blocks will likely get orphaned, so your concern is unfounded.
However, by the time the blocks are consistently big, the technology will already have improved to be able to handle them.
With static limit, big blocks now will become small blocks in the future and the process will have to repeat itself.

This is a good way to keep the system in check, allow it grow, while at the same time not let it spiral out of our hands.

Why exactly will they get orphaned? I consider this statement unfounded, because nowadays miners are able to propagate blocks between themselves in constant time (generally) with help of Matt Corallo's relay network. In turn, the nodes don't have this privilege, so we better have some testing, I guess.

[coalitionfor8mb]

In fact, the size of the block is not limited by either core devs or miners, but ISPs around the world. I believe without major upgrade of internet infrastructure, the bitcoin network can not handle 2MB blocks well, not even mention 8MB

When there was a talk about 20MB blocks, people calculated that they would barely fit into the "total bandwidth per month cap" imposed by ISPs.
So, I assumed that 8MB should fit that situation better, provided that it is only a cap and not the actual block size for the next few years.

Why exactly will they get orphaned? I consider this statement unfounded, because nowadays miners are able to propagate blocks between themselves in constant time (generally) with help of Matt Corallo's relay network. In turn, the nodes don't have this privilege, so we better have some testing, I guess.

Ok, there are finally some good points here!
We definitely need some testing.

But if the "constant propagation time" for miners is true, why Chinese are then worried about the bandwidth with their 8MB figure?
What would prevent implementing the relay network for other full nodes and what was the rationale for not doing so in the first place?

[RoadTrain]

Why exactly will they get orphaned? I consider this statement unfounded, because nowadays miners are able to propagate blocks between themselves in constant time (generally) with help of Matt Corallo's relay network. In turn, the nodes don't have this privilege, so we better have some testing, I guess.

Ok, there are finally some good points here!
We definitely need some testing.

But if the "constant propagation time" for miners is true, why Chinese are then worried about the bandwidth with their 8MB figure?
What would prevent implementing the relay network for other full nodes and what was the rationale for not doing so in the first place?

The thing about the relay network is that it depends on miners behaving cooperatively. On the other hand, some game-theoretical research (e.g. selfish mining paper) suggests that this is not necessarily the most rational approach.
So, my point is that we must be very careful about changing these constants. I'd rather go with 2MB first, not 8MB, and then re-evaluate.

[coalitionfor8mb]

Why exactly will they get orphaned? I consider this statement unfounded, because nowadays miners are able to propagate blocks between themselves in constant time (generally) with help of Matt Corallo's relay network. In turn, the nodes don't have this privilege, so we better have some testing, I guess.

Ok, there are finally some good points here!
We definitely need some testing.

But if the "constant propagation time" for miners is true, why Chinese are then worried about the bandwidth with their 8MB figure?
What would prevent implementing the relay network for other full nodes and what was the rationale for not doing so in the first place?

The thing about the relay network is that it depends on miners behaving cooperatively. On the other hand, some game-theoretical research (e.g. selfish mining paper) suggests that this is not necessarily the most rational approach.
So, my point is that we must be very careful about changing these constants. I'd rather go with 2MB first, not 8MB, and then re-evaluate.

I also thought about more conservative approach first, but then quickly realized that being over-protective would necessitate having this discussion (about the limit increase) more often than we want it to be.

In the extreme case, we will be constantly debating the issue and not let the actual limit to solidify in order for it to have any effect, which in turn would distract us from doing any other positive and creative things in the ecosystem.

If 8MB is considered unsafe, then we should look at how competition is currently surviving with higher limits (not blocks though). Of course, the counter argument in this case will be that their market cap and importance are less than that of Bitcoin and therefore there are no incentives to disturb the network by storming it with bigger blocks.

It's still somewhat uncertain, but if we demonstrate that we can agree on a single number and move forward, then the positive momentum in the ecosystem might encourage other players (like bandwidth-limited Chinese miners) to protect the network from bloating too fast by building on top of blocks that are not bigger than 2MB, for example.

That's the scenario I keep in mind when talking about 8MB limit.