Mempool continues to grow could this be the tipping point?

[FUBAR-BDHR]

It's going to start getting really interesting if the mempool continues to grow like this into the halving. With an already growing mempool even if low percentage of miners turn off their machines the resulting longer blocks until the next difficulty could send the mempool skyrocketing out of control. If it does will there be a panic sell forcing even more miners to turn off equipment?


Right now mempool is close to 29meg that's 5 hours of blocks (assuming average 10 minutes) to clear it even if no new transactions are added.

[LiteCoinGuy]

last year the mempool was at 55MB without a problem. it would take 1-2 GB of data to crash a node. i think we are in the green zone at the moment.

[iFixBTCmemoryIssues]

It's going to start getting really interesting if the mempool continues to grow like this into the halving. With an already growing mempool even if low percentage of miners turn off their machines the resulting longer blocks until the next difficulty could send the mempool skyrocketing out of control. If it does will there be a panic sell forcing even more miners to turn off equipment?


Right now mempool is close to 29meg that's 5 hours of blocks (assuming average 10 minutes) to clear it even if no new transactions are added.

We'll be good when everyone starts using solutions that dosen't hold Bitcoin back.

Core and their direction is already a disaster that is escalating by the day.

"My transaction is delayed", "My fee is wrong", "The fees are too high" - I hear this on a daily basis from dozens of people. Now its, "the mempool is out of control, my node could crash". Great.

A bunch of guys pushing a solution on us that we don't want, we don't need and that is costing us money.

[FUBAR-BDHR]

last year the mempool was at 55MB without a problem. it would take 1-2 GB of data to crash a node. i think we are in the green zone at the moment.

How many new people won't use bitcoin again? How many existing people are pissed off and leaving? How many businesses are being hurt because transactions can't get through?

[iFixBTCmemoryIssues]

last year the mempool was at 55MB without a problem. it would take 1-2 GB of data to crash a node. i think we are in the green zone at the moment.

How many new people won't use bitcoin again? How many existing people are pissed off and leaving? How many businesses are being hurt because transactions can't get through?

My clients businesses of whom take payments in Bitcoin have furious customers.

I just got off the phone with one of my clients who said to me, "Nobody is going to use Bitcoin".

[arnoudk]

last year the mempool was at 55MB without a problem. it would take 1-2 GB of data to crash a node. i think we are in the green zone at the moment.

This very much depends on your definition of green zone.

If you mean that nodes won't technically crash at the same time all around the world - then yes.

Most nodes have a 300 MB mempool config, so that should be sufficient.

However, if you are talking about user experience, we are in the orange zone. If you are new to bitcoin and are making your first transaction - you really cannot be expected to know exactly what fee to pay. Selecting 'normal' on the wallet should be enough.

The Default behavior of wallet software should give a Great user experience.

When I made my first bitcoin transaction, I was worried I had done perhaps something wrong. It was a little stressful until the funds arrived where they needed to go. And that was when bitcoin was working the way it should. If it takes many hours to confirm (and you've been told bitcoin is fast!) then I would have been wondering what I was doing wrong. Did I lose my money? Where was it, and could I ever get it back? Damn, there is no company to call. No customer service number. No recourse.

You need to do a few good bitcoin transactions to trust the system. (I only started to understand REALLY how it worked after sending some money). If that trust is damaged due to a bad first experience, I don't think we can talk about being in the green zone.

I think you are strictly looking at the technical state of the system, and from that perspective it has not collapsed. Nor do I think that it will. Transactions will get dropped, but nodes won't crash. Given that definition, bitcoin will be in the green zone even if it tanks to zero.

[arnoudk]

Now its, "the mempool is out of control, my node could crash". Great.

Wasn't this problem solved?

I remember that Bitcoin XT had a very simple (but effective) mempool eviction strategy that basically dropped a random transaction from the mempool - as a preferred strategy to crashing and thus dropping all transactions from memory.

Before that, it USED to be possible to fill up the mempool, as it did not have a maximum size, and have the bitcoin application use more memory than the computer physically has. When that happens, bitcoin crashes with an Out of Memory Exception, as would be expected.

The flood of bitcoin transactions back then caused bitcoin to evolve, and become more anti-fragile. It strengthened and that immunity is saving the day today!

I don't believe that crashing bitcoin through mempool flooding is still possible. Do you have any evidence to suggest otherwise? I believe all main implementations have limits on the size of the memory, including BU, Classic, XT and even Core.

[arnoudk]

My clients businesses of whom take payments in Bitcoin have furious customers.

I just got off the phone with one of my clients who said to me, "Nobody is going to use Bitcoin".

Yes, there are a few things that really piss me off about this whole situation.
1. User experience for people using bitcoin is degraded. If you know what bitcoin can do (you have used it before), you may be able to accept a lapse in performance more easily than if you just used it to send your first payment.
2. Using bitcoin for the first time can be scary (when sending a non-trivial amount of money for you). You need a quick network confirmation to grow your confidence. Monies should arrive in their destination wallets as expected. There will be stress when using a new payment network that you do not understand for the first time. That stress should be limited.
3. The default behavior of wallet software - regardless of which one you use - should give a good initial experience. There is a disconnect between network performance and wallet functionality.
4. I really cringe when I read the 'you suck because you screwed up you moron' replies on reddit. Telling people new to bitcoin that they screwed up because they did not check some obscure third party fee calculation site that they could not have known about as a new user, and telling them to 'just wait 3 days for the transaction to be canceled due to their own stupidity' is completely unacceptable. Since when is it OK to kick new members of this community in the face while laughing at them at the same time?!?
5. If you as a business accept bitcoin, YOU are going to be the one that customers call for support (who else!). Poor bitcoin network performance, or poor user experience (whatever the cause) reflects on the business accepting bitcoin - whether it is their fault or not. This increases the barrier to acceptance for businesses (that, and increased support costs for customer queries).
6. It was so predictable that this problem was going to occur, so long ago. It was completely avoidable. That we are in this situation, while KNOWING about it for years, reflects very poorly on bitcoin.

Yet, I can't believe that miners will continue to make the wrong choice for much longer. They can't afford to. Would they really ALL throw away their entire investment in state of the art chips, and get stuck with killing power utility contracts - rather than choosing what their customers clearly demand? That makes no sense, and I still think they will do the right thing.

Historically, people are lazy. Even when facing certain attack - they rather hope the problem will go away. However, when things go south - actions are fierce, decisive and much stronger than they needed to be. History repeats itself. It would be better to have a measured response to full blocks, but I think we will have a slightly-too-late and very strong reaction to correct. Hopefully on time.

[arnoudk]

How many new people won't use bitcoin again? How many existing people are pissed off and leaving? How many businesses are being hurt because transactions can't get through?

You don't get a second chance for a first impression - and right now the first impression is bad for far too many people.

For them, it does not matter what causes it. All that matters, is that they had an expectation going into bitcoin - and an experience that does not match.

[inBitweTrust]

There is a lot of ignorance and misinformation around. The default mempool set for nodes is 300MB and half of that is for txs. This has been the case for a very long time. So the statistic of "29 MB " mempool isn't very useful and is just fear mongering.

What we should be disusing is the atrocity that some wallets still don't dynamically use floating fees. We need to educate users to avoid using wallets like Jaxx and blockchain.info. It is also fine to discuss whether you think that 8-9 pennies for an average tx size getting confirmed in 0-2 blocks is an appropriate fee when the network is has a backlog of txs. Whether this is a spam attack or natural increase in txs the question is still the same. Is 4-5 pennies per tx under low load and 8-9 pennies under high load a reasonable amount to spend ?

[iFixBTCmemoryIssues]

Now its, "the mempool is out of control, my node could crash". Great.

Wasn't this problem solved?

I remember that Bitcoin XT had a very simple (but effective) mempool eviction strategy that basically dropped a random transaction from the mempool - as a preferred strategy to crashing and thus dropping all transactions from memory.

Before that, it USED to be possible to fill up the mempool, as it did not have a maximum size, and have the bitcoin application use more memory than the computer physically has. When that happens, bitcoin crashes with an Out of Memory Exception, as would be expected.

The flood of bitcoin transactions back then caused bitcoin to evolve, and become more anti-fragile. It strengthened and that immunity is saving the day today!

I don't believe that crashing bitcoin through mempool flooding is still possible. Do you have any evidence to suggest otherwise? I believe all main implementations have limits on the size of the memory, including BU, Classic, XT and even Core.

For nodes running the latest releases, yes.

XT designed an extremely smart feature. Something I admire about all productive implementations is that they each bring something different to the table.

Interestingly enough, when the mempool was originally overloaded, I wrote a script to dump the transactions. It worked as a hack job. I believe I posted that or sent it in a PM to someone as they were having the issue. There was little information back then and so I spent a few days fixing it myself.

[iFixBTCmemoryIssues]

My clients businesses of whom take payments in Bitcoin have furious customers.

I just got off the phone with one of my clients who said to me, "Nobody is going to use Bitcoin".

Yes, there are a few things that really piss me off about this whole situation.
1. User experience for people using bitcoin is degraded. If you know what bitcoin can do (you have used it before), you may be able to accept a lapse in performance more easily than if you just used it to send your first payment.
2. Using bitcoin for the first time can be scary (when sending a non-trivial amount of money for you). You need a quick network confirmation to grow your confidence. Monies should arrive in their destination wallets as expected. There will be stress when using a new payment network that you do not understand for the first time. That stress should be limited.
3. The default behavior of wallet software - regardless of which one you use - should give a good initial experience. There is a disconnect between network performance and wallet functionality.
4. I really cringe when I read the 'you suck because you screwed up you moron' replies on reddit. Telling people new to bitcoin that they screwed up because they did not check some obscure third party fee calculation site that they could not have known about as a new user, and telling them to 'just wait 3 days for the transaction to be canceled due to their own stupidity' is completely unacceptable. Since when is it OK to kick new members of this community in the face while laughing at them at the same time?!?
5. If you as a business accept bitcoin, YOU are going to be the one that customers call for support (who else!). Poor bitcoin network performance, or poor user experience (whatever the cause) reflects on the business accepting bitcoin - whether it is their fault or not. This increases the barrier to acceptance for businesses (that, and increased support costs for customer queries).
6. It was so predictable that this problem was going to occur, so long ago. It was completely avoidable. That we are in this situation, while KNOWING about it for years, reflects very poorly on bitcoin.

Yet, I can't believe that miners will continue to make the wrong choice for much longer. They can't afford to. Would they really ALL throw away their entire investment in state of the art chips, and get stuck with killing power utility contracts - rather than choosing what their customers clearly demand? That makes no sense, and I still think they will do the right thing.

Historically, people are lazy. Even when facing certain attack - they rather hope the problem will go away. However, when things go south - actions are fierce, decisive and much stronger than they needed to be. History repeats itself. It would be better to have a measured response to full blocks, but I think we will have a slightly-too-late and very strong reaction to correct. Hopefully on time.

Your entire post is exactly how I see things and am experiencing them. From new users to businesses and even the community "frustrated" and "angry" as to whats going on in their Bitcoin. We see Core saying, "if we don't get our way, we will drop the price to zero". They can't do this, at worst they will drop it into the 300s which I am prepared for.

On this forum I say a lot about, "Core is headed for destruction" and what you wrote explains that. They don't talk to the businesses, or the customers, only the ones on their side with vested interests. Even if they do now, if you can't win me (I use to run Core for all my businesses and clients), you can't win them. I am a very easy-going person to work with, just don't steal from me.

Core will be around for the next half a decade with suffering numbers, adoption and then in the end, they will be sitting around saying to each other - "what did we do wrong?".

This just isn't my feeling, its your feeling, Rogers feeling, every person I speak to in business and ask their opinion has the same feeling. New businesses that are integrating Bitcoin and Ethereum.

I can explain what you did wrong, instead of working with us to help you - you went on a crazy mission of destruction.

[iFixBTCmemoryIssues]

There is a lot of ignorance and misinformation around. The default mempool set for nodes is 300MB and half of that is for txs. This has been the case for a very long time. So the statistic of "29 MB " mempool isn't very useful and is just fear mongering.

What we should be disusing is the atrocity that some wallets still don't dynamically use floating fees. We need to educate users to avoid using wallets like Jaxx and blockchain.info. It is also fine to discuss whether you think that 8-9 pennies for an average tx size getting confirmed in 0-2 blocks is an appropriate fee when the network is has a backlog of txs. Whether this is a spam attack or natural increase in txs the question is still the same. Is 4-5 pennies per tx under low load and 8-9 pennies under high load a reasonable amount to spend ?

Keep thinking that and blaming wallet software, blame everyone but yourselves.

When blockchain.info sees your post, I hope they make the right business decision and throw your Lightning network under the bus. I would never tell customers to avoid my clients business as I would be working with them to help in every way possible.

I do get a good chuckle reading your posts, it continues to confirm your destruction.

When your game is over, see below:

[arnoudk]

There is a lot of ignorance and misinformation around. The default mempool set for nodes is 300MB and half of that is for txs. This has been the case for a very long time. So the statistic of "29 MB " mempool isn't very useful and is just fear mongering.

What we should be disusing is the atrocity that some wallets still don't dynamically use floating fees. We need to educate users to avoid using wallets like Jaxx and blockchain.info. It is also fine to discuss whether you think that 8-9 pennies for an average tx size getting confirmed in 0-2 blocks is an appropriate fee when the network is has a backlog of txs. Whether this is a spam attack or natural increase in txs the question is still the same. Is 4-5 pennies per tx under low load and 8-9 pennies under high load a reasonable amount to spend ?

I agree that some technical metric such as the number of megabytes of a mempool is not relevant for the end user.

A full mempool - if caused by fee paying transactions - impacts user experience however. It is an indication for long delays, and for high pricing. It does not necessarily guarantee those situations exist, but it is a good indicator for that.

Indeed, nodes no longer crash when the mempool becomes full. During one of the previous peaks (I think that was during the so called stress-test) it became obvious that a max size needed to be defined. Bitcoin has become more robust because of it and that anti-fragility is helping it now. Yes, I know that the default mempool setting is 300 MB (I mentioned that in one of my posts above), I am not worried about technical failure of the network.

The metric that is relevant, is how long it takes for transactions to confirm (in minutes) on average when the bitcoin blocks are at capacity. It is also relevant what the fees are.

Sending a proper fee, is the job of a good wallet. Right now, too many wallets still fail at this. It is really important that the default setting on a wallet gives a good, and affordable, user experience all of the time. This is not so easy - fee estimation in a dynamic environment involves a lot of guessing.

Ensuring that bitcoin can grow in the user base, while keeping transaction pricing competitive, is the job of the consensus rules (block size).

BOTH need to improve today.

I accept Antonopoulos's definition that any transaction that pays an appropriate fee, is a genuine transaction. Any other definition would be arbitrary and central-planned one. I believe that bitcoin needs to be able to make a simple transaction (one or two inputs, one or two outputs) for just a few cents. 2c, 3c maybe. It should not be more than this, else bitcoin will be the domain of wealthy westerners, while it has the potential to serve billions of un- or underbanked. They cannot afford tens of cents for each transaction, and will pay it only if they have no choice. Bitcoin cannot use coercion, so it needs to be fiercely competitive.

[inBitweTrust]

It is an indication for long delays, and for high pricing.

I make txs daily and always get in the next block. I think once I had to wait till the 2nd block. Your statement is only accurate if people use poorly designed wallets or are purposely setting the fees low.

I accept Antonopoulos's definition that any transaction that pays an appropriate fee, is a genuine transaction. Any other definition would be arbitrary and central-planned one.

I'm fine with this definition , as long as the sender doesn't externalize the costs of "spamming" the network by inserting viruses on the block chain, putting penis hash graffiti, avoid using sendmany properly , sending tx back in forth with a do while loop to attack the network, ect ... The plan always has been for fees to slowly start subsidizing security rather than new users doing this through inflation. This is exactly what is happening here. Block reward dropping and fees going up! I understand that you would likely rather have many more txs cover the costs or the fees instead; the numbers do not add up and this cannot scale however. I would much rather have tx fees to be 10 dollars eventually on the chain and 1-3 pennies on LN nodes, this can scale and will insure a secure and healthy ecosystem.

Any other definition would be arbitrary and central-planned one. I believe that bitcoin needs to be able to make a simple transaction (one or two inputs, one or two outputs) for just a few cents. 2c, 3c maybe. It should not be more than this, else bitcoin will be the domain of wealthy westerners, while it has the potential to serve billions of un- or underbanked. .

Sure , this can happen with the LN and payment channels. In the meantime it is better that fees are growing to secure the network.
We are headed in the right direction.

[iFixBTCmemoryIssues]

It is an indication for long delays, and for high pricing.

I make txs daily and always get in the next block. I think once I had to wait till the 2nd block. Your statement is only accurate if people use poorly designed wallets or are purposely setting the fees low.

I accept Antonopoulos's definition that any transaction that pays an appropriate fee, is a genuine transaction. Any other definition would be arbitrary and central-planned one.

I'm fine with this definition , as long as the sender doesn't externalize the costs of "spamming" the network by inserting viruses on the block chain, putting penis hash graffiti, avoid using sendmany properly , sending tx back in forth with a do while loop to attack the network, ect ... The plan always has been for fees to slowly start subsidizing security rather than new users doing this through inflation. This is exactly what is happening here. Block reward dropping and fees going up! I understand that you would likely rather have many more txs cover the costs or the fees instead; the numbers do not add up and this cannot scale however. I would much rather have tx fees to be 10 dollars eventually on the chain and 1-3 pennies on LN nodes, this can scale and will insure a secure and healthy ecosystem.

Any other definition would be arbitrary and central-planned one. I believe that bitcoin needs to be able to make a simple transaction (one or two inputs, one or two outputs) for just a few cents. 2c, 3c maybe. It should not be more than this, else bitcoin will be the domain of wealthy westerners, while it has the potential to serve billions of un- or underbanked. .

Sure , this can happen with the LN and payment channels. In the meantime it is better that fees are growing to secure the network.
We are headed in the right direction.

Miners will make more from the block-reward this year than they did last year (Roger Ver posted a graph).

As for blaming wallets or users, this has nothing to do with them. This has to do with an organized scam and lies you peddle.

We're headed in the correct direction with Bitcoin Unlimited, these guys work hard and have no financial motive besides the network performing well.

Time will tell

[inBitweTrust]

We're headed in the correct direction with Bitcoin Unlimited, these guys work hard and have no financial motive besides the network performing well.

Time will tell

I sincerely do hope that you are successful with the BU implementation. I don't look at this as an us vs them battle and am not interested in trolling and nasty inflammatory remarks to divide the community. Any proven and secure solution is ok in my book and If BU proves itself and comes out with some good testing than great.* Good luck friend and lets move forward.


* Xthin blocks vs Compact Blocks is a good example of two competing solutions . May the best one be used , but thank goodness we are testing multiple solutions and BU is digging up old solutions to retest.

[iFixBTCmemoryIssues]

We're headed in the correct direction with Bitcoin Unlimited, these guys work hard and have no financial motive besides the network performing well.

Time will tell

I sincerely do hope that you are successful with the BU implementation. I don't look at this as an us vs them battle and am not interested in trolling and nasty inflammatory remarks to divide the community. Any proven and secure solution is ok in my book and If BU proves itself and comes out with some good testing than great.* Good luck friend and lets move forward.


* Xthin blocks vs Compact Blocks is a good example of two competing solutions . May the best one be used , but thank goodness we are testing multiple solutions and BU is digging up old solutions to retest.

I am not your friend, I do not associate with terrorists who attack my business.

A word from the wise, you are in over your head.

[arnoudk]

I make txs daily and always get in the next block. I think once I had to wait till the 2nd block. Your statement is only accurate if people use poorly designed wallets or are purposely setting the fees low.

Yes, if you pay more than your competition, you will get in the next block or two. However, the question is - what happens if everyone starts paying higher fees? What of course will happen is that the most valuable transactions (to their senders) get confirmed, and the less valuable ones never make it into a block. The rich can use bitcoin. The poor cannot. The west can use bitcoin. The third world can not. Banks, or settlement providers van use bitcoin. Regular consumers can not. Etc.

Don't confuse your experience with the experience of everyone. I don't doubt that your experience has been good. I have never had to wait longer than an hour myself, and I still think that is acceptable. Yet I know others who are having more problems.

Also don't underestimate how much more bitcoin savvy you are than the normal users. You understand what 'satoshi per bytes' is and can price your transaction accordingly. I can do that, too - but we are a minority. I think this works BECAUSE we are still a minority.

If many of the transactions are unimportant to their owners, they will get priced out first. This is why, in the past, with stress testing, no real users were impacted. At some point, most transactions are real world people sending real transactions to each other. Maybe that time is now, maybe tomorrow, maybe a month or two from now. Who knows. But once these real transactions start competing for a place in the block, prices rise quickly to exclude some people from using bitcoin. Like I said before, the third world unbanked population will be excluded.

Yes, wallet design needs to be improved. No argument there.

But space in the blocks also needs to be increased.

I accept Antonopoulos's definition that any transaction that pays an appropriate fee, is a genuine transaction. Any other definition would be arbitrary and central-planned one.

I'm fine with this definition , as long as the sender doesn't externalize the costs of "spamming" the network by inserting viruses on the block chain, putting penis hash graffiti, avoid using sendmany properly , sending tx back in forth with a do while loop to attack the network, ect ... The plan always has been for fees to slowly start subsidizing security rather than new users doing this through inflation. This is exactly what is happening here. Block reward dropping and fees going up! I understand that you would likely rather have many more txs cover the costs or the fees instead; the numbers do not add up and this cannot scale however. I would much rather have tx fees to be 10 dollars eventually on the chain and 1-3 pennies on LN nodes, this can scale and will insure a secure and healthy ecosystem.

Ok, so you are not fine with this definition. Because you add a whole lot of extra clauses after the 'as long as' criteria. I understand the sentiment somewhat, but the problem is that it is arbitrary. Why can I not put penis hash graffiti on the blockchain (actually, I would be surprised if it were not already there) if I pay the appropriate fee? Why can I put a hash of some document in there, but not ASCII art? If it is important enough for me to pay the fee, I do not consider it a problem. In fact, the more fee paying transactions, the better. Disk storage is cheap (I don't suppose you subscribe to the idea that the storage of the blockchain on disk is an issue?). So storing ASCII art pays for the security of the system.

I fully accept that transaction fees will and must replace the block reward. However, let's do some simple maths. Lets say a block can hold 2000 transactions and the block reward is 25 BTC. That means that, to maintain a 25 BTC income per block, each transaction would have to pay 0.0125 BTC. That is 10 dollars. (I assume you calculated the same thing). I can pay 10 dollars for an international wire. But not for a domestic one. The number of use cases where I am prepared to pay 10 dollar are, in fact, so limited that I would use it maybe once a year. (But I probably wouldn't use it, because there will most likely be a cheaper alternative). I also wouldn't have any bitcoin to send, because nobody would like to spend 10 dollars to pay me (or, put another way, I'd rather have the extra 10 dollars in my pocket). Since I wouldn't have bitcoin, and there is no real demand for bitcoin to buy anything useful, I don't think a foreign party will like receiving bitcoin. All they could do with it, is send it to the other side of the world for their payment - but couldn't realistically use it to pay their office rent or employee salary.

However, if the transactions were to cost 2-5c, I would use it multiple times per day. I'd pay 40 dollars per year in tx costs, more or less, to use bitcoin. That is 4x more income for miners. Not profit of course, but income.

I don't think data traffic, data storage or transaction verification is a bottleneck - now or in the near future. I think it is easier to earn the same amount of money from transaction fees when there are more transactions, than when you require people to pay so much.

Sure , this can happen with the LN and payment channels. In the meantime it is better that fees are growing to secure the network.

Oh sure. But this requires that LN and payment channels and side chains etc etc actually exist, are implemented, and can be used.

I would love for these second layer networks to exist on top of bitcoin, and I would love for them to make transacting more affordable. They certainly have a place in the ecosystem. If transacting via LN is cheaper than transacting via the main blockchain - people will voluntarily choose to use LN when that makes sense. You don't need to discourage the use of the main block chain artificially, the fees will do that for you.

You can increase the block size. Remember that you are increasing the max size, not the actual size. If there are not many transactions (because LN has been rolled out, is in use, and is efficient) the block size will probably decrease - even if you increase the max block size.

Transactions will move, from the expensive chain to the cheap LN.

But, right now, LN does not exist. (It is in development, but it is not deployed and from the perspective of the users it does not yet exist). And, right now, people are being excluded from bitcoin. The people who could benefit most, the poor people in third world countries, are priced out of the system. Sure, for most western people, 10 dollars is nothing. Here in Belize, that is a DAYS work for an average employee. In some parts of Africa, I believe this is one or more weeks work. Would you work for a day, to be able to pay the transaction fee of a single transaction??

The answer to that is, you probably would, if you must send money and have no alternative. If you need to do an international wire and can choose between western union and bitcoin - you'll choose bitcoin. But you are not going to request payment of your salary in bitcoin. You'd resist that fiercely, because you can't afford to spend your money.

Try not to look at bitcoin only from a western perspective. The west is not where bitcoin has the most potential.

Earning 2c on a billion transactions a day is 120 bitcoin per block in transaction fees. If everyone on the planet sent one transaction per week - you'd be there.

Really, the profit is in a massive quantity of cheap transactions, not a handful of overpriced ones.

We are headed in the right direction.

Yes, we are.

[inBitweTrust]

Yes, if you pay more than your competition, you will get in the next block or two. However, the question is - what happens if everyone starts paying higher fees? What of course will happen is that the most valuable transactions (to their senders) get confirmed, and the less valuable ones never make it into a block. The rich can use bitcoin. The poor cannot. The west can use bitcoin. The third world can not. Banks, or settlement providers van use bitcoin. Regular consumers can not. Etc.

While this is correct in principle it has 2 incorrect assumptions in practice:
1) It ignores that fact that there are currently many spam txs on the network thus there is indeed capacity if we bump just out price these txs -
See this for examples of what is spam -
https://en.bitcoin.it/wiki/Flood_attack
https://en.bitcoin.it/wiki/Spam_transactions

2) It assumes that I and others don't also want to raise capacity as needed and no , we dont think high tx fees is a good thing for consumers. I would suggest anything above 30 pennies is too high per tx because that is the territory where it starts becoming the same price merchant processors charge for cc's to merchants for small txs. I believe we can do better with payment channels and bring fees down to 2-3 pennies though and offchain could be free.

Don't confuse your experience with the experience of everyone. I don't doubt that your experience has been good. I have never had to wait longer than an hour myself, and I still think that is acceptable. Yet I know others who are having more problems.

This is another point I disagree upon . Waiting even 5 seconds for a confirmation is completely unacceptable for in person payments. This is why we need layer 2 solutions , and kicking the can down the road with a blocksize bump removes the incentive to start using these.

Also don't underestimate how much more bitcoin savvy you are than the normal users. You understand what 'satoshi per bytes' is and can price your transaction accordingly. I can do that, too - but we are a minority. I think this works BECAUSE we are still a minority.

which is why we need to steer new users aways from horrible wallets like jaxx and blockchain.info and toward decent wallets like mycelium, copay, and core that use floating fees that do all the math for the consumer . There is absolutely no reason we need to go to https://bitcoinfees.21.co/ and manually calculate fees . I only use that to investigate and troubleshoot why others txs are getting stuck and than I find out they use one of these horrible wallets.

But space in the blocks also needs to be increased.

Agree , segwit will give us around 1.86 Mb , schnorr signatures a little more capacity but we will indeed need a HF eventually.

[inBitweTrust]

Because you add a whole lot of extra clauses after the 'as long as' criteria. I understand the sentiment somewhat, but the problem is that it is arbitrary. Why can I not put penis hash graffiti on the blockchain (actually, I would be surprised if it were not already there) if I pay the appropriate fee? Why can I put a hash of some document in there, but not ASCII art? If it is important enough for me to pay the fee, I do not consider it a problem. In fact, the more fee paying transactions, the better. Disk storage is cheap (I don't suppose you subscribe to the idea that the storage of the blockchain on disk is an issue?). So storing ASCII art pays for the security of the system.

I agree that much of spam is arbitrary and subjective (not flood attacks though , which is very deliberate spam) and have already suggested that this is all fine and we should allow spam on the network- even flood attacks..... as long as the costs aren't externalized and the sender is paying the fee in a fee market instead of getting new users to subsidize their flood attacks and penis pictures. You have to understand that such spam not only burdens the network temporarily but permanently!

I fully accept that transaction fees will and must replace the block reward. However, let's do some simple maths. Lets say a block can hold 2000 transactions and the block reward is 25 BTC. That means that, to maintain a 25 BTC income per block, each transaction would have to pay 0.0125 BTC. That is 10 dollars. (I assume you calculated the same thing). I can pay 10 dollars for an international wire. But not for a domestic one. The number of use cases where I am prepared to pay 10 dollar are, in fact, so limited that I would use it maybe once a year. (But I probably wouldn't use it, because there will most likely be a cheaper alternative). I also wouldn't have any bitcoin to send, because nobody would like to spend 10 dollars to pay me (or, put another way, I'd rather have the extra 10 dollars in my pocket). Since I wouldn't have bitcoin, and there is no real demand for bitcoin to buy anything useful, I don't think a foreign party will like receiving bitcoin. All they could do with it, is send it to the other side of the world for their payment - but couldn't realistically use it to pay their office rent or employee salary.

Scaling on chain is a non starter when you do the math.

On chain -
7 billion people making 2 blockchain transactions per day
24 GB blocks (~2,400 MB)
7e9 * 2txs * 250Bytes / 144 block/day
~50 Mbit/s best-case with IBLT @ 2tx/day/person


Lightning Network
7 billion people roll-over 2 channels per year
133 MB blocks - unlimited transactions count
7e9 * 2 txs * 500Byte / 52560blocks/yr
~3 Mbit/s with IBLT

[arnoudk]

While this is correct in principle it has 2 incorrect assumptions in practice:
1) It ignores that fact that there are currently many spam txs on the network thus there is indeed capacity if we bump just out price these txs -
See this for examples of what is spam -
https://en.bitcoin.it/wiki/Flood_attack
https://en.bitcoin.it/wiki/Spam_transactions

I just don't accept the 'this transaction pays the appropriate fee, but I STILL consider it spam because <insert arbitrary reason that may change tomorrow>.

Transactions that do not pay enough fee are already deprioritized and are already not included in blocks (even if there is space).

What you are calling spam transactions, are transactions that have enough fee. They are important enough for someone to actually fork over that money to include it in a block. Not you, nor I, have anything to say about that in a trustless, decentralized peer to peer system. By claiming your opinion of spam is valid, you are undermining one of the most sacred core principles of bitcoin.

If the block size would increase, it would be great for the miners. They get all this extra income processing transactions. Because transaction fees are part of the miner income, these transactions that you call spam are strengthening the network.

You can't have it both ways. You can't say that you want to have transaction fees replace the block reward, and also say you want to limit paying transactions by some arbitrary and changing spam definition.

2) It assumes that I and others don't also want to raise capacity as needed and no , we dont think high tx fees is a good thing for consumers. I would suggest anything above 30 pennies is too high per tx because that is the territory where it starts becoming the same price merchant processors charge for cc's to merchants for small txs. I believe we can do better with payment channels and bring fees down to 2-3 pennies though and offchain could be free.

Well, do you want to raise capacity today? Do you support 2 MB blocks right now? Are you in favor of a hard fork to make this happen - implemented today and activated in the near future?

If the answer to that is No, then please don't pretend you want to raise the block size.

30c per transaction is still ridiculous, but your math goes completely out of the window with this figure. 2000 tx * 0.30 dollar per tx = 600 dollar per block. At today's exchange rate, that is 600 / 750 = 0.8 BTC per block. That is quite a haircut you are advocating for the miners - from 25 right now to 0.8.

If you wanted to keep 25 bitcoins as a block reward after the halving, you would need to have enough transactions for 12.5 BTC (to make up the difference). $0.30 = 0.0004 BTC tx fee per transaction. So, 12.5/0.0004 = 31250 tx per block. Assuming a 1 MB block holds 2000 tx, this is 31250/2000 = 15.63 MB blocks.

Segwit provides, MAYBE a 1.8 MB block size. So your argument appears to break down from what I can tell. Where am I misunderstanding you?

Payment channels might be cheaper, but they do not exist right now. You just cannot solve an immediate emergency with the promise of a possible solution in the future. That is a very dangerous logical fallacy...

Don't confuse your experience with the experience of everyone. I don't doubt that your experience has been good. I have never had to wait longer than an hour myself, and I still think that is acceptable. Yet I know others who are having more problems.

This is another point I disagree upon . Waiting even 5 seconds for a confirmation is completely unacceptable for in person payments. This is why we need layer 2 solutions , and kicking the can down the road with a blocksize bump removes the incentive to start using these.

Ok, we can disagree on that. The bitcoin I signed up for did not have immediate confirmations, but it worked good enough for accepting transactions that were unconfirmed. RBF complicates that, because now you need to examine if an incoming transaction is RBF, and if so, you cannot accept that payment...

If layer 2 solutions require artificially crippling bitcoin, then these layer 2 solutions suck!

Really, they should be able to compete on their own merit. They should be faster, cheaper, have instant verifications, etc. Their added value is what should make them desirable. They should compete on a level playing field. Not in an environment where the competitor is prevented to grow and service these people due to a centrally planned limit.

Now don't get me wrong. I love these layer 2 solutions. I want them to exist. I want them to compete with on chain activity and ensure that the transactions happen in those chains that makes most sense for the sender. I think LN has a great potential, but it does not compete fully with all bitcoin transactions.

"Fiat - a currency so good that it's use has to be mandatory."
"Ligntning Network - a layer 2 solution so good that it's use has to be mandatory".

LN should compete, in a free and fair and level market. I welcome it with open arms in that scenario. But I will shun it, completely, if it is forced down my throat with artificial limitations on bitcoin. Because, obviously, then it would be a worse solution.

which is why we need to steer new users aways from horrible wallets like jaxx and blockchain.info and toward decent wallets like mycelium, copay, and core that use floating fees that do all the math for the consumer . There is absolutely no reason we need to go to https://bitcoinfees.21.co/ and manually calculate fees . I only use that to investigate and troubleshoot why others txs are getting stuck and than I find out they use one of these horrible wallets.

Sure, I already agreed with you that wallets need to improve. But I made it quite clear that this is only one part of the problem.

But space in the blocks also needs to be increased.

Agree , segwit will give us around 1.86 Mb , schnorr signatures a little more capacity but we will indeed need a HF eventually.

[/quote]
So, do you want 1.8MB blocks? (I hope 1.6 megabit was a typo!).

Or do you want 30c transactions requiring 15.63 MB blocks?

Or do you want miners to earn significantly less?

You just can't have it all. I am really sorry - but either you are talking about 10 dollars per transaction with 1 MB blocks, or you are talking 30c per transaction with 16 MB blocks.

[arnoudk]

I agree that much of spam is arbitrary and subjective (not flood attacks though , which is very deliberate spam) and have already suggested that this is all fine and we should allow spam on the network- even flood attacks..... as long as the costs aren't externalized and the sender is paying the fee in a fee market instead of getting new users to subsidize their flood attacks and penis pictures. You have to understand that such spam not only burdens the network temporarily but permanently!

Well, if there is enough room in the blocks, and the fee is high enough, there IS no flood attack. It simply does not exist. Just process the transactions, take their money, and be happy. You are arguing a point here that exists only because you want to limit block sizes. If not, please speak out that you support bigger block sizes to be implemented Right Now, as in today.

Scaling on chain is a non starter when you do the math.

You did not respond to my math. Does that mean that you accept it, but are only concerned with the block sizes? Does that mean that if I can alleviate your fears for the size of a block - that you are fully on board? I'd prefer to keep the discussion focused by first agreeing to something before throwing new variables in the mix.

On chain -
7 billion people making 2 blockchain transactions per day
24 GB blocks (~2,400 MB)
7e9 * 2txs * 250Bytes / 144 block/day
~50 Mbit/s best-case with IBLT @ 2tx/day/person

Ok, so the ENTIRE world is making TWO block chain transactions every single day. You and I would be so rich at this point, because we were able to buy bitcoin at $750 a piece today.

It is extremely unlikely that this scenario will occur within the next years. So, we have a lot of improvements in technology, storage, network speed, processing speed, etc to go. In this scenario, there will be ASICs doing the signature validation, and these ASICs will be cheap and readily available. Your CPU would not have to do all this work. In fact, if every person on the planet used bitcoin, these chips would be embedded in PCs.

But, even if that were to happen TODAY - are you saying that 50 Mbps is an issue? (Actually, my maths comes to 40 MBps: 24000 MB blocks / 600 seconds per interval = 40 MB per second).

Let's compare this to Usenet.
https://en.wikipedia.org/wiki/Usenet

Usenet is not anywhere near as important as a global payments network that has replaced all banks, credit cards, and money transmitters (as in your example). Yet it uses 23870000 MB per DAY. That is 276 MB per second.

For 276 MB per second, you get a distributed news platform that is used mainly to share binaries, such as HD videos and pictures of cats.

Explain to me why 40 MB per second, almost 7x less traffic, would be an issue for a global financial network?

I really do not care if some teenager is able to run bitcoin from his bedroom on an ADSL line he shares with mum, dad and his sister.

Lightning Network
7 billion people roll-over 2 channels per year
133 MB blocks - unlimited transactions count
7e9 * 2 txs * 500Byte / 52560blocks/yr
~3 Mbit/s with IBLT

Wishful thinking.

I am sorry, but this is completely and utterly nonsense, if you are claiming that the third world will be using Lightning Network for their payments. It proves to me that you have not spent a single second observing how unbanked people live, and how they use money.

You are once again looking at this from an ivory tower of western wealth. What you are suggesting, is completely incompatible with third world reality.

Let me give you an example.

When I pay someone at the end of the day for a day's work, he takes that fiat currency and puts it in his pocket. (He does not have a wallet, nor does he need one). He immediately goes to buy food for THAT DAY, buy whatever necessities he must have that he can afford now, and possibly pay off some debt (everyone owes everyone here). The money is completely gone within the hour. When he goes to bed, he is completely broke.

Now, you are suggesting, he puts enough money in a LN payment channel, so that he needs only 2 per YEAR? I am sorry, but that just shows you do not understand how poor people live.

So, you are excluding the third world. You are excluding the billions of people that bitcoin could help. And I don't understand why.... is it because some teenager needs to be able to run bitcoin on his Atari?

[LiteCoinGuy]

back to normal.

[FUBAR-BDHR]

And it started to go right back up earlier until a string of fast hitting blocks brought it back down again. The only reason it dropped yesterday was it was Sunday and transaction load slows down almost every Sunday.

14 blocks in the last hour.

[LiteCoinGuy]

And it started to go right back up earlier until a string of fast hitting blocks brought it back down again. The only reason it dropped yesterday was it was Sunday and transaction load slows down almost every Sunday.

14 blocks in the last hour.

can you post a longer chart?

[arnoudk]

back to normal.

Yes, but this does not mean that all transactions were processed that were in the queue. They expire after 3 days. Part of the decline is undoubtedly caused by blocks being mined. I suspect part of the decline is also expiration after 3 days.

If I look at my disk storage of unconfirmed transactions (I save the mempool to mysql), I see over 35.000 unconfirmed transactions. I can't easily see how many are fee paying vs free riders... so it is not necesarilly an indication that transactions are being missed. But still it is a lot higher than I have seen in a long while.

[iFixBTCmemoryIssues]

I agree that much of spam is arbitrary and subjective (not flood attacks though , which is very deliberate spam) and have already suggested that this is all fine and we should allow spam on the network- even flood attacks..... as long as the costs aren't externalized and the sender is paying the fee in a fee market instead of getting new users to subsidize their flood attacks and penis pictures. You have to understand that such spam not only burdens the network temporarily but permanently!

Well, if there is enough room in the blocks, and the fee is high enough, there IS no flood attack. It simply does not exist. Just process the transactions, take their money, and be happy. You are arguing a point here that exists only because you want to limit block sizes. If not, please speak out that you support bigger block sizes to be implemented Right Now, as in today.

Scaling on chain is a non starter when you do the math.

You did not respond to my math. Does that mean that you accept it, but are only concerned with the block sizes? Does that mean that if I can alleviate your fears for the size of a block - that you are fully on board? I'd prefer to keep the discussion focused by first agreeing to something before throwing new variables in the mix.

On chain -
7 billion people making 2 blockchain transactions per day
24 GB blocks (~2,400 MB)
7e9 * 2txs * 250Bytes / 144 block/day
~50 Mbit/s best-case with IBLT @ 2tx/day/person

Ok, so the ENTIRE world is making TWO block chain transactions every single day. You and I would be so rich at this point, because we were able to buy bitcoin at $750 a piece today.

It is extremely unlikely that this scenario will occur within the next years. So, we have a lot of improvements in technology, storage, network speed, processing speed, etc to go. In this scenario, there will be ASICs doing the signature validation, and these ASICs will be cheap and readily available. Your CPU would not have to do all this work. In fact, if every person on the planet used bitcoin, these chips would be embedded in PCs.

But, even if that were to happen TODAY - are you saying that 50 Mbps is an issue? (Actually, my maths comes to 40 MBps: 24000 MB blocks / 600 seconds per interval = 40 MB per second).

Let's compare this to Usenet.
https://en.wikipedia.org/wiki/Usenet

Usenet is not anywhere near as important as a global payments network that has replaced all banks, credit cards, and money transmitters (as in your example). Yet it uses 23870000 MB per DAY. That is 276 MB per second.

For 276 MB per second, you get a distributed news platform that is used mainly to share binaries, such as HD videos and pictures of cats.

Explain to me why 40 MB per second, almost 7x less traffic, would be an issue for a global financial network?

I really do not care if some teenager is able to run bitcoin from his bedroom on an ADSL line he shares with mum, dad and his sister.

Lightning Network
7 billion people roll-over 2 channels per year
133 MB blocks - unlimited transactions count
7e9 * 2 txs * 500Byte / 52560blocks/yr
~3 Mbit/s with IBLT

Wishful thinking.

I am sorry, but this is completely and utterly nonsense, if you are claiming that the third world will be using Lightning Network for their payments. It proves to me that you have not spent a single second observing how unbanked people live, and how they use money.

You are once again looking at this from an ivory tower of western wealth. What you are suggesting, is completely incompatible with third world reality.

Let me give you an example.

When I pay someone at the end of the day for a day's work, he takes that fiat currency and puts it in his pocket. (He does not have a wallet, nor does he need one). He immediately goes to buy food for THAT DAY, buy whatever necessities he must have that he can afford now, and possibly pay off some debt (everyone owes everyone here). The money is completely gone within the hour. When he goes to bed, he is completely broke.

Now, you are suggesting, he puts enough money in a LN payment channel, so that he needs only 2 per YEAR? I am sorry, but that just shows you do not understand how poor people live.

So, you are excluding the third world. You are excluding the billions of people that bitcoin could help. And I don't understand why.... is it because some teenager needs to be able to run bitcoin on his Atari?

Your post is brilliant!!!!!!!!!!!!!!!!

I am in this for the billions of people, that's how we will make real change + it will benefit us greatly!!!!!!!!!!!!!!

I can wait for the trillions, I am in no rush

[rogerver]

Is 4-5 pennies per tx under low load and 8-9 pennies under high load a reasonable amount to spend ?

No, this will drive people to use something else.