Will the blockchain transform into an inter-cluster-settlement system?

[allocater]

Premise: It is not sustainable that billions of users send millions of transaction through the block chain every second. It will just grow too much.

if wrong, goto end and post reply.

if true, continue:

So we need to put for example 10% of users into cluster A, 10% into cluster B, .... 10% into cluster J.

Now all millions of transactions are collected for a while (10min?) and all transactions within clusters are settled in the cluster internally. Only if there are settlements to be done between clusters, those go into the blockchain. The blockchain now only contains few giant sum transactions with giant fees that go from cluster A to B, or C to G, J to D, etc. that represent these inter-cluster-settlements.

Congratulations you have just eroded the bitcoin system and established the old banking system again. Cluster = bank. Inter-cluster settlements = inter-bank settlements (aka Target2)   

What to do?

Maybe some kind of dynamic fractal clustering, where each cluster, subcluster, subsubcluster,... has it's own blockchain and every time something can not be resolved within the subsubsubsubcluster blockchain it gets transfered to a higher cluster blockchain, bitcoins becomes thousands of blockchains and you can 'subscribe' to the subsubsubsubcluster you like, or something. The blockchain we know now, becomes the inter-continental blockchain, that only will contain transaction/settlements between continents, or Planets, or Solar Systems.

[justusranvier]

Premise: It is not sustainable that billions of users send millions of transaction through the block chain every second. It will just grow too much.

Instead of just accepting that premise, why not instead ask, "What changes would need to happen for Bitcoin to allow 10 billion people to each generate 100 transactions per day by 2030?"

That's an upper limit of about 1 million transactions per second.

[misterbigg]

ask, "What changes would need to happen for Bitcoin to allow 10 billion people to each generate 100 transactions per day by 2030?"

Due to the broadcast nature of network and the way that storage grows in proportion to O(M*N) where M = number of transactions and N = number of nodes, it doesn't seem likely that Bitcoin in its current incarnation (regardless of cosmetic changes) will scale to anything near that level.

[justusranvier]

ask, "What changes would need to happen for Bitcoin to allow 10 billion people to each generate 100 transactions per day by 2030?"

Due to the broadcast nature of network and the way that storage grows in proportion to O(M*N) where M = number of transactions and N = number of nodes, it doesn't seem likely that Bitcoin in its current incarnation (regardless of cosmetic changes) will scale to anything near that level.

Serious question - why did you even bother to post that?

[misterbigg]

Serious question - why did you even bother to post that?

Well your numbers were...lets call them "large"    but the meme that "Bitcoin should scale to handle a high volume of microtransactions" is rampant in here and I think it's worth pointing out that handling that volume is not really practical in the Bitcoin paradigm.

[hashman]

Congratulations you have just eroded the bitcoin system and established the old banking system again. Cluster = bank. Inter-cluster settlements = inter-bank settlements (aka Target2)   

Not quite.  These "clusters" or third party payment system overlay providers still can't create an infinite amount of coins.  Further, we could check to block chain to see how much they are holding.  But yeah, just because there are bitcoins out there doesn't mean people are going to instantly be cured of their mental illnesses. 

[justusranvier]

Well your numbers were...lets call them "large"     but the meme that "Bitcoin should scale to handle a high volume of microtransactions" is rampant in here and I think it's worth pointing out that handling that volume is not really practical in the Bitcoin paradigm.

It adds zero value to the discussion to point out that the existing implementation can not scale to the necessary values. Everybody knows that already. Currently Bitcoin utilizes bandwidth and storage space in the least efficient way possible. Of course it can't continue to do that and scale to a million transactions per second. Nobody is suggesting that it can, so stop pointing out the obvious as if you are bringing something useful to the table.

[misterbigg]

...Everybody knows that already...Nobody is suggesting that it can, so stop pointing out the obvious as if you are bringing something useful to the table.

Everybody does not know that already, which is apparent every time the subject of block limit comes up. Plenty of people who don't understand the nature of the protocol assume that simply removing the limit will allow the number of transactions handled per second to rise to arbitrary levels. But I don't see why there is the need for hostility.

[ShadowOfHarbringer]

ask, "What changes would need to happen for Bitcoin to allow 10 billion people to each generate 100 transactions per day by 2030?"

Due to the broadcast nature of network and the way that storage grows in proportion to O(M*N) where M = number of transactions and N = number of nodes, it doesn't seem likely that Bitcoin in its current incarnation (regardless of cosmetic changes) will scale to anything near that level.

Then lets make it scale to that level.

We can, for example, make a **cluster client**, which will scale to multiple machines and supercomputers, just like Linux kernel scales. It will be run on supercomputers, graphics cards, and mini-clusters like beowulf.

Possibilities are endless. So please stop talking jibberish, because you are not making any sense.

[justusranvier]

But I don't see why there is the need for hostility.

Because you're being disingenuous. Despite having this pointed out multiple times in countless threads you are still propagating falsehoods.

"Bitcoin can not scale to high transaction rates on affordable hardware" is false.

"Bitcoin requires improvements to the way it stores and broadcasts blocks in order to scale to high transaction rates on affordable hardware" is true.

[ShadowOfHarbringer]

But I don't see why there is the need for hostility.

Because you're being disingenuous. Despite having this pointed out multiple times in countless threads you are still propagating falsehoods.

I think he has a hidden agenda (which is not pro-bitcoin).
Or maybe he is just some stupid kid.

[misterbigg]

I think he has a hidden agenda (which is not pro-bitcoin). Or maybe he is just some stupid kid.

I'm neither anti-Bitcoin nor am I young (my avatar is a photo of me). I have significant Bitcoin exposure, and I'm quite happy with it. Maybe I need to re-read what's been said about scaling up. This having been said, there's no need to be an asshole about it.

[ShadowOfHarbringer]

I think he has a hidden agenda (which is not pro-bitcoin). Or maybe he is just some stupid kid.

I'm neither anti-Bitcoin nor am I young (my avatar is a photo of me). I have significant Bitcoin exposure, and I'm quite happy with it. Maybe I need to re-read what's been said about scaling up. This having been said, there's no need to be an asshole about it.

If you don't want me to be an asshole, first stop being an asshole yourself.

It has been alread explained multiple times that your arguments addressing scalability of Bitcoin (and also some other) are completely and utterly wrong, yet you keep pushing your opinion trying to convince everybody that you are right even than you are not.

Stop spreading your bullshit around like a hippo with diarrhea, and i will stop being an asshole. OK ?

[tralala]

But I don't see why there is the need for hostility.

Because you're being disingenuous. Despite having this pointed out multiple times in countless threads you are still propagating falsehoods.

"Bitcoin can not scale to high transaction rates on affordable hardware" is false.

"Bitcoin requires improvements to the way it stores and broadcasts blocks in order to scale to high transaction rates on affordable hardware" is true.

And in what way should that be possible?
Please give concrete suggestions.

edit: I've logged just 7 hours so far reading this forum so it's likely I haven't seen the posts in which it is explained how you can make Bitcoin scale to 1000+ transactions/s, so please be a bit patient with me if I don't know the suggested ideas already.

[arklan]

But I don't see why there is the need for hostility.

Because you're being disingenuous. Despite having this pointed out multiple times in countless threads you are still propagating falsehoods.

"Bitcoin can not scale to high transaction rates on affordable hardware" is false.

"Bitcoin requires improvements to the way it stores and broadcasts blocks in order to scale to high transaction rates on affordable hardware" is true.

And in what way should that be possible?
Please give concrete suggestions.

edit: I've logged just 7 hours so far reading this forum so it's likely I haven't seen the posts in which it is explained how you can make Bitcoin scale to 1000+ transactions/s, so please be a bit patient with me if I don't know the suggested ideas already.

the simplest explanation i am aware of is two fold: one, altering the blocksize limit so allow more then 7 transactions a second at maximum. whether this is done, how it is done, when it is done, is a major discussion point.

of course, in addition to allowing more transactions, we need a way to handle the ever increasing size of the blockchain. this is where pruning comes in. i'm not totally familiar with the technical aspects of it, but it basically boils down to not needing to keep of record of transactions that have been entirely spent in the past. if/when implemented, this will dramatically reduce the size of the stored blockchain. combined with "light" clients that don't store it, like web wallets and such, that should resolve this particular issue.

[justusranvier]

https://bitcointalk.org/index.php?topic=159014.0

[chriswen]

Instead of "clusters" people can use different currencies that will promote transfer speeds, ex litecoin.  Until litecoin becomes a value currency too.

[tvbcof]

I say knuckle down and make the current solution rock solid and retain the ability for many millions of private individuals to operate the system without putting their life savings at risk by creating the  compute clusters needed to be a peer.

What I would suspect will naturally fall out will be a unassailable value core with a multitude of loosely coupled end-user solutions (alternate currencies, on-line wallets, Joe operating a node for his group of friends, etc.)  Trying to stifle such a solution will be an impossible whack-a-mole task.

I will again bet almost all of my BTC (literally) that when the solution tries to scale to service the economic activity that is now almost certain, and likely to appear faster than development can safely proceed, Bitcoin will become so centralized and specialized that it will be easy to cut the head off.

[tralala]

Thanks for the links and the repsonses.

Increasing the blocksize limit is a possible short- to midterm solution and it's increasingly discussed so this might be undertaken in the near future. I imagine a blocksize of 10 MB is possible thus increasing the limit to 70tr/s. This is already quite impressive and should be enough for a couple of years.

But:

Bigger blocksize does not only increase the storage requirements but also the time other miners verify a block and thus the risk of producing an orphan block. I've not yet heard of a solution how this might be circumvented.

I think the longterm solution IF Bitcoin becomes a transaction currency instead of a speculation object is more off-blockchain transactions. How to make those secure and reliably integrated in the main blockchain will be a challenge.

[justusranvier]

Increasing the blocksize limit is a possible short- to midterm solution and it's increasingly discussed so this might be undertaken in the near future. I imagine a blocksize of 10 MB is possible thus increasing the limit to 70tr/s. This is already quite impressive and should be enough for a couple of years.

Gavin has all but said an increase will happen in his last development update.

Bigger blocksize does not only increase the storage requirements but also the time other miners verify a block and thus the risk of producing an orphan block. I've not yet heard of a solution how this might be circumvented.

That's a feature, not a bug.

https://bitcointalk.org/index.php?topic=140233.msg1503099#msg1503099

I think the longterm solution IF Bitcoin becomes a transaction currency instead of a speculation object is more off-blockchain transactions. How to make those secure and reliably integrated in the main blockchain will be a challenge.

Off-blockchain transactions are a terrible idea, and aren't even necessary. The optimizations needed to overcome the challenges you are referring to are not a mystery.