My thoughts about the blocksize thing

[Gabi]

Since everyone and his dog is openin thread about that, i'll hop on the bandwagon too 

But my doubt is another one, right now, with the 1MB limit, the blockchain at max can grow of like 55GB per year. If we have the necessity to increase the block size, this mean that every block is already 1MB and that in a year the blockchain will be like 60GB.
If we double the blocksize, it will be 110GB per year. In 4 years it will be like 440GB. With the current limit, 220GB, still a lot

I am not against increasing the blocksize, if we have lot of transactions, it must be done. Also more transactions per block means more fees for the miners, after all it is not expensive to make a 1mb or a 2mb block, but a 2mb block will have more transactions. Yes, the 1mb block would have less transactions but with higher fees but i think the 2mb block in total would have more fees.

And as i said, having a blockchain of 220 or 440gb is not a big difference, it is a huge chain anyway. The problem is that it is huge, even with the current limit. A bit too huge.

[JackH]

Thing is new tech can solve this and we wont need to have the same debate: http://www.thinlinedata.com/dna-storage-harddrives/

But then again, Bitcoin is ahead of its time, unless some synthetic biologist moves his *** 

[prezbo]

Also don't forget that the majority of the blockchain can be pruned when it becomes too large.

[Gabi]

Wich is exactly what i was thinking. But so far there is no real work done about that. If we want to speak about increasing the blocksize, then pruning must be considered too.

[justusranvier]

But so far there is no real work done about that. If we want to speak about increasing the blocksize, then pruning must be considered too.

Depends on your definition of "work".

[kokojie]

For regular clients, we don't really have to store old transactions, who really care about a transaction happened 5 years ago? for those who care, you can have the option of storing the
full chain. For the rest of us, we'll just store recent transactions. So for example, on installation of the client, you can have the option of:

* store recent transactions
* store full chain

For those who want to look up a transaction 10 years ago, blockchain.info or blockexplorer.com etc... can provide that service by storing the full chain.

[markm]

One of the other threads started out with exactly the concern that increasing the block size before all the pruning and so on are even being used just makes less incentive to actually do the damn pruning and so on and so on.

Until all those other things are in use, the space and bandwidth we already have/use is not being efficiently used so throwing good space after bad is just silly.

When the size we already have is being used in the most efficient way, which by the way involves bandwidth stuff as well as disk space utilisation stuff, we can look then at whether, given all those miracle cures that the moar size moar size moar size crowd point to as solutions to the problem of moar size but still have not actually put into place are in place, moar size is actually even needed.

I wonder what the average block size, not counting free transactions, is for the last difficulty-adjustment period?

-MarkM-

[misterbigg]

i'll hop on the bandwagon too...more transactions per block means more fees for the miners

While you're hopping on the bandwagon you might to...actually read the posts in the Development & Technical Discussion forum regarding block size instead of merely regurgitating your view that "hard coded limits are bad." More transactions per block means less fees for miners. I'll leave it as an exercise for you to figure out why.

[prezbo]

Wich is exactly what i was thinking. But so far there is no real work done about that. If we want to speak about increasing the blocksize, then pruning must be considered too.

I absolutely agree.

[conv3rsion]

i'll hop on the bandwagon too...more transactions per block means more fees for the miners

While you're hopping on the bandwagon you might to...actually read the posts in the Development & Technical Discussion forum regarding block size instead of merely regurgitating your view that "hard coded limits are bad." More transactions per block means less fees for miners. I'll leave it as an exercise for you to figure out why.

If this is true, and miners have the option currently of including fewer (or 0) transactions in their blocks, why do current blocks not have minimal numbers of transactions?

It seems to me that to maximize transaction revenue, a miner would want to include the max number of paying transactions that he could, and a higher number per block would equal higher fees.

[conv3rsion]

I don't understand why some people want a hard limit to encourage fewer potential transactions. Fewer transactions equals higher fees per transactions and less useful technology (at least in the role of a payment network) which then equals a less valuable and used technology (in all other roles).

A raised max block limit that increases at a slower rate than corresponding increases in commodity hardware and bandwidth does not mean that the majority of people wishing to act as full nodes cannot do so. The original 1MB limit is completely arbitrary and should scale as all other resources scale. The only reason that 21,000,000 total coins is not considered a problem is because of divisibility. Blocks also need to be continually divisible for Bitcoin to be valuable. Taking that ability away is incredibly shortsighted.

[Technomage]

There are other ways to go around the fact that the blockchain becomes massive. Not everyone needs to have the whole blockchain of all Bitcoin existence. Why wouldn't the last 6 months be quite sufficient, for example, and maybe have the rest as SPV? Or something like that. We could have separate archive nodes that just store the whole thing.

It isn't needed that all full nodes store all transactions ever done. Only the ones done in recent history are critical.

[Gabi]

More transactions per block means less fees for miners

I do not agree.
More transactions with lower fee per transaction can give in total more fees than less transactions with higher fees.

[twolifeinexile]

I don't understand why some people want a hard limit to encourage fewer potential transactions. Fewer transactions equals higher fees per transactions and less useful technology (at least in the role of a payment network) which then equals a less valuable and used technology (in all other roles).

A raised max block limit that increases at a slower rate than corresponding increases in commodity hardware and bandwidth does not mean that the majority of people wishing to act as full nodes cannot do so. The original 1MB limit is completely arbitrary and should scale as all other resources scale. The only reason that 21,000,000 total coins is not considered a problem is because of divisibility. Blocks also need to be continually divisible for Bitcoin to be valuable. Taking that ability away is incredibly shortsighted.

I am not sure this is so much about the hard limit rather than the unintended consequences if hard limit is raised and removed all together.  These wide-spread discussion is very good in itself.

Also important is the stability of the protocol, if you can raise the block size today, then how about total bitcoin tommorrow? We all know they are different, how and where is the line? And who determine the change? Can people opt to resist the change easily?

All these should be well-thought and discussed.

[prezbo]

More transactions per block means less fees for miners

I do not agree.
More transactions with lower fee per transaction can give in total more fees than less transactions with higher fees.

And let's not forget that miners are the ones setting the required fees. If it's not profitable for them to include free transactions, they just won't.

[Akka]

And let's not forget that miners are the ones setting the required fees. If it's not profitable for them to include free transactions, they just won't.

Sorry, but I can't sai this often enough. Long term, Mining wont be providable. No matter how high the fees are. Simply due to the fact that as long as you can make money with mining there will always be new Hashpower added until it reaches a equilibrium, where it won't be providable any more to buy mining equipment.

As soon as direct buyable Mining equipment is available this will go very fast. The free marked dictates this.

The amount of money made by mining will only set how much money is invested to secure the network.

Mining is (soon) a near zero sum game.

[Jaw3bmasters]

There are other ways to go around the fact that the blockchain becomes massive. Not everyone needs to have the whole blockchain of all Bitcoin existence. Why wouldn't the last 6 months be quite sufficient, for example, and maybe have the rest as SPV? Or something like that. We could have separate archive nodes that just store the whole thing.

It isn't needed that all full nodes store all transactions ever done. Only the ones done in recent history are critical.

I thought a version was already available that doesn't download the entire chain, just recent transactions........

[markm]

Here is a thought I have not seen around yet:

Litecoin probably already provides four megabytes of block per ten minutes. Is it the lack of mention of that specific "improvement over bitcoin" in their marketing that accounts for the failure of the moar moar moar crowd to abandon the pathetically small megabytes of block per ten minutes bitcoin for the four times "better" (in this oh so very important we'd guess from the number of threads about it) "problem" of bitcoin's?



-MarkM-

P.S. Check out Liquidcoin too! Spew out as many blocks as you like as fast as you like!

[conv3rsion]

I don't understand why some people want a hard limit to encourage fewer potential transactions. Fewer transactions equals higher fees per transactions and less useful technology (at least in the role of a payment network) which then equals a less valuable and used technology (in all other roles).

A raised max block limit that increases at a slower rate than corresponding increases in commodity hardware and bandwidth does not mean that the majority of people wishing to act as full nodes cannot do so. The original 1MB limit is completely arbitrary and should scale as all other resources scale. The only reason that 21,000,000 total coins is not considered a problem is because of divisibility. Blocks also need to be continually divisible for Bitcoin to be valuable. Taking that ability away is incredibly shortsighted.

I am not sure this is so much about the hard limit rather than the unintended consequences if hard limit is raised and removed all together.  These wide-spread discussion is very good in itself.

Also important is the stability of the protocol, if you can raise the block size today, then how about total bitcoin tommorrow? We all know they are different, how and where is the line? And who determine the change? Can people opt to resist the change easily?

All these should be well-thought and discussed.

Nobody (that I have seen) is advocating for the limit to be removed all together, they are advocating for it to scale along with realistic resource availability and consumption. Why is the block limit 1MB and not 10KB? Because 1MB was seen as having realistic resource availability 4 years ago and it was anticipated to be easy to grow in the future. Too bad Satoshi's 1st version wasn't absolutely perfect from the beginning, I guess we are forever stuck with a shitty wire replacement system that will eventually lose value and be replaced since ideologues can't be bothered to give 2 fucks about actual adoption.  

Limiting the number of coins is not a problem because they are divisible. Limiting the number of "possible" transactions forever is a problem because it is counter the primary goal of any technology, which is to be used.

[prezbo]

Sorry, but I can't sai this often enough. Long term, Mining wont be providable.

I don't necessarily agree. It's obvious if the state of bitcoin remains like it is now, when the reward drops to zero security by mining will drop ridiculously low. There are only two ways to make it work (keeping the same security), either increase the fees or increase the number of transactions (with fees) per block. There are currently about 0.5 btc of transaction fees per block, so if the transaction rate would increase 40-fold that would take care of things.