At any given point in time the entire BTC networks txs are handled by 1 miner

[notig]

If you ignore the aspects of bitcoin about security and such.. and just focus on the transaction part of the network... it appears that at any given point in time the transaction part of bitcoin is essentially always done by just 1 miner.

If you ignore mining pools and just imagine a bunch of individual miners hoping to get a hash lower than the target this is easy to see. Joining a mining pool doesn't somehow magically make the transactions have more capability with all that added power... it just guarantees you a sort of return rather than relying on the sheer luck of mining.

I don't mean to say this is a problem.... this is a scalability issue.. and it's probably unavoidable.

But you should keep this in mind when you talk about the block size limit because that is a scalability issue on top of a scalability issue. Now not only is the entire networks transactions being essentially handled by 1 miner at any given point of time... but the block size limits how much that one miner can include in his block.

With the progress of technology and the fact that we are doing okay now it may not seem like a dire situation. But the block size fixed at 1MB means we will not ever reap the benefits of improved technology and we will always be stuck at a level that is already below what the majority of people are probably capable of. That's just a guess.

Correct me if I'm wrong(or if I am misunderstanding). I probably am wrong but I'd like to know exactly how.... I have more reading to do .

[DeathAndTaxes]

Well it is more accurate to say all transactions are processed independently by all miners (unless a miner chooses to exclude on from the current block).

Still I am not sure what scalability limit you are concerned about.  The current codebase is capable of validating in excess of 4,000 tps on an average CPU.  When you consider that PayPal transaction rate is ~50 tps and the US fedwire system ($663 trillion USD in annual transaction volume) is about 8 tps it kinda puts that into perspective.

[Monster Tent]

If you ignore the aspects of bitcoin about security and such.. and just focus on the transaction part of the network... it appears that at any given point in time the transaction part of bitcoin is essentially always done by just 1 miner.

If you ignore mining pools and just imagine a bunch of individual miners hoping to get a hash lower than the target this is easy to see. Joining a mining pool doesn't somehow magically make the transactions have more capability with all that added power... it just guarantees you a sort of return rather than relying on the sheer luck of mining.

I don't mean to say this is a problem.... this is a scalability issue.. and it's probably unavoidable.

But you should keep this in mind when you talk about the block size limit because that is a scalability issue on top of a scalability issue. Now not only is the entire networks transactions being essentially handled by 1 miner at any given point of time... but the block size limits how much that one miner can include in his block.

With the progress of technology and the fact that we are doing okay now it may not seem like a dire situation. But the block size fixed at 1MB means we will not ever reap the benefits of improved technology and we will always be stuck at a level that is already below what the majority of people are probably capable of. That's just a guess.

Correct me if I'm wrong(or if I am misunderstanding). I probably am wrong but I'd like to know exactly how.... I have more reading to do .

If certain bitcoin sites improved their efficiency the block size issue is a non issue. But why do that when you can spread FUD instead.

[justusranvier]

If certain bitcoin sites improved their efficiency the block size issue is a non issue. But why do that when you can spread FUD instead.

Sooner or later Bitcoin will need to have the additional transaction capacity, even if the one badly-behaving site you're talking about wasn't around.

Looking at the attention Bitcoin is getting in the media, and the rate at which new sites are adoption Bitcoin for payment, and the rate at which exchanges and payment processors are starting up around the world, my bet is on "sooner".

[DannyHamilton]

Well it is more accurate to say all transactions are processed independently by all miners (unless a miner chooses to exclude on from the current block).

Still I am not sure what scalability limit you are concerned about.  The current codebase is capable of validating in excess of 4,000 tps on an average CPU.  When you consider that PayPal transaction rate is ~50 tps and the US fedwire system ($663 trillion USD in annual transaction volume) is about 8 tps it kinda puts that into perspective.

Current codebase has a limit of 1MB per block.
Typical transaction (1 input, 2 outputs) is about 260 bytes

1048576 bytes / 260 bytes = 4033 transactions per block.

Block occurs on average every 10 minutes (600 seconds)

4033 transactions every 600 seconds = 4033 / 600 = 6.7 tps.

Current codebase is capable of confirming 6.7 tps (less than PayPal and US fedwire).

Is my math wrong?

Perhaps a 7.5 MB max blocksize would be a good idea since it would provide enough space to eventually compete with PayPal?

[DeathAndTaxes]

Yes I understand the block limit issue but the OP seemed to indicate there is a " scalability issue on top of a scalability issue".  A mining node with solid specs and decent connectivity presents no bottleneck at any reasonable transaction speed.  


I will stay out of the hard limit debate simply because I think it is premature but given even with the 1M limit Bitcoin can operate at close to the transaction speed of FedWire, I think this blockchain debate is a tempest in a teacup.  I mean lets be conservative and say that someday Bitcoin can "only" support half a quadrillion dollars in value transacted annually, and the rest go to lesser altchains or off blockchain transactions.  Oh no what a collasal failure this half a quadrillion dollar network is. 

My take, lets get to the point where Bitcoin is transactions trillions upon trillions of dollars worth of transactions on network with hundreds of thousands of full nodes and tens of millions of end users before we worry about the "hard limit".  At this point it is like a shipbuilder telling Columbus "why try and build faster ships nothing can go faster than the hard limit called the speed of light?"

[DannyHamilton]

Yes I understand the block limit issue but the OP seemed to indicate there is a " scalability issue on top of a scalability issue". . .

Yeah, I wasn't paying close enough attention.  You're right the OP seems to be indicating that there is a scalability issue before the blocksize.  I agree with you, I don't see it.

[niko]

Yes I understand the block limit issue but the OP seemed to indicate there is a " scalability issue on top of a scalability issue". . .

Yeah, I wasn't paying close enough attention.  You're right the OP seems to be indicating that there is a scalability issue before the blocksize.  I agree with you, I don't see it.

Me neither. It is worth pointing out that the transactions are not part of what is being repeatedly hashed (with increasing nonce) while mining, only the block header. There is no scalability issue there: hitting the target is equally probable for a block with one or one hundred transactions.
OP: is this correct?

[markm]

Damn, forgot about the speed of light. We're doomed whatever we do, at least the pony express produced fertilizer, not sure these discussions are all that fertile anymore...

How could God have been so short-sighted? A hard limit, srsly? What was she thinking?

-MarkM-

[ShadowOfHarbringer]

Perhaps a 7.5 MB max blocksize would be a good idea since it would provide enough space to eventually compete with PayPal?

No constant blocksize will ever be good, because
1. Market, economy and people are dynamic things.
2. Changing the limit every time we need to is unfeasible. It is better to solve things once and for all instead of changing it frequently to cope with increased needs.
3. 1MB was just a temporary solution for BETA period. Even satoshi mentioned that the limit will have to be changed some time in the future.

IMO blocksize should always scale with the network/demand, the same way as difficulty scales.
So there should be either some blockchain voting process, an automated algorithm which decides proper block limit or even both.

[oleganza]

IMO blocksize should always scale with the network/demand, the same way as difficulty scales.
So there should be either some blockchain voting process, an automated algorithm which decides proper block limit or even both.

Block size will scale with network/demand completely automatically when there is no artificial limit. So far the limit does not matter and blocks grow according to the amount of transactions (automatically!). No need to invent any regulatory algorithm to keep the limit in place, it won't matter anyway. The first moment 90% of people touch the limit, it will be raised without asking your opinion. So, effectively, there will be no limit, just some number in the source code as a memorial to FUD around the topic.

I wrote already why this will happen: https://bitcointalk.org/index.php?topic=146918.msg1559226#msg1559226

[DannyHamilton]

Perhaps a 7.5 MB max blocksize would be a good idea since it would provide enough space to eventually compete with PayPal?

No constant blocksize will ever be good, because
1. Market, economy and people are dynamic things.
2. Changing the limit every time we need to is unfeasible. It is better to solve things once and for all instead of changing it frequently to cope with increased needs.
3. 1MB was just a temporary solution for BETA period. Even satoshi mentioned that the limit will have to be changed some time in the future.

IMO blocksize should always scale with the network/demand, the same way as difficulty scales.
So there should be either some blockchain voting process, an automated algorithm which decides proper block limit or even both.

I really wish someone would invent a sarcasm font for the internet.

[ShadowOfHarbringer]

The first moment 90% of people touch the limit, it will be raised without asking your opinion.

Nope it won't as long as 95% of clients will not relay such blocks.
Miners cannot raise the limit on their own without the official client supporting all sizes of blocks.

Also, i think that some additional code will be needed in order to auto-reject too large blocks, but I may be wrong.

I really wish someone would invent a sarcasm font for the internet.

Oh well, i didn't read your post completely - just few lines, so no way of detecting the sarcasm. Sorry.

[edd]

Damn, forgot about the speed of light.
...

How could God have been so short-sighted? A hard limit, srsly? What was she thinking?

-MarkM-

Priceless!

[Peter Todd]

Regarding the maximum transaction rate, I worked it out using the most efficient possible transaction type and wrote up the results on the wiki.

Basically, 10.7tx/s is possible, although the actual rate will be somewhere between that and 5.2tx/s depending on how efficiently change can used managed.