Bitcoin Forum

Get ready, it's coming...

http://gavinandresen.ninja/time-to-roll-out-bigger-blocks

For what purpose is the bigger block? Why is there a need for a 1 megabyte block, the blockchain works fine like now?
Blocks only get big when there are many transaction in them, so how could a miner create a bigger block without that much transactions

Okay. This may be a stupid question, but I'm going to ask anyway. Are bigger blocks good or bad? And why would they be good or bad?

It's better to be proactive than reactive, there's a ton of discussion about the "why" already out there.

About time! Anyone whining about this need to get their marbles examined.

increase block size may injured bitcoin node ... because of the pollution of trading (high trading).
people that need speed ... must build sidechain (or offchain strategy like bitpay ou coinbase).

i don't agree to increase block size except if it has a limitation ... (like 10MB).

That makes sense. Thank you for answering my question.

The blockchain size is already turning into a serious problem...  Don't speak with such a vicious tone when you don't know what you are talking about.

When the time comes, if we want to take over the world (that's having the volume of transactions of say VISA, or all of them combined -mastercard, paypal, etc etc- we'll need a bigger block size to not saturate the blockchain. The problem is, as far as im concerned it would require a fork, so 2 Bitcoins would exist if some of the stuck up idiot purists don't abandon the 1MB blockchain. Correct me if im wrong.

except the size was already ok before gavin changed it once already, now it's being changed again and will need to be again down the road instead of allowing the network to decide (miners) he's taken it upon himself to set hardcoded limits where there was none originally.

but my question is why they choose 20mb as the next size? 5 or 10 weren't enough? did he tested those value?

I think that's right. It could be 5 or 10, but we don't want to have to do this every year.

The real block size is up to the miners. The protocol (wired in the bitcoin reference implementation) can suggest a MAX_SIZE and miners can agree on that. However, they can mine blocks with size between [MIN_SIZE, MAX_SIZE] and that means they can include any number of transactions per block always they respect the agreed MAX_SIZE. Note that it is possible to mine blocks with zero transactions for example.

IMO there is an incentive to keep the block size small because it makes the space in them scarce, and we know what happen when something is scarce. Given the block size is 1MB, if your transaction requires 60kb you have to pay more fee than if your transaction is 800bytes. So, bigger blocks could reduce the fees but at the same time miner are allowed to introduce more transactions per block, what means more fees for them.

I really don't fully understand all the implications around this change because even when the current block MAX_SIZE is 1 MB, miners don't mine 1MB blocks. In fact, I read that one of the most common sizes is around 731kb. Clearly, for miners is not the same to broadcast to the network 1MB than 20MB because if I mine a 20MB block and, at the same time, other miner mines a 200Kb block, its block will spread faster than the mine one.

So, I don't see the incentive for miners now. What I mean is that currently they earn more with the 12.5 bitcoins that the protocol gives them that with the transactions fee and for that reason I don't know if more transactions per block (more fees) worth the risk.
      

Of course it isn't.  If you don't want to store the full chain, run an SPV client.  If you do: http://www.amazon.com/Seagate-Desktop-3-5-Inch-Internal-ST4000DM000 (http://www.amazon.com/Seagate-Desktop-3-5-Inch-Internal-ST4000DM000).  Pruning is coming, so the storage "issue" will soon vanish completely.

Increasing the max block size does not mean that all blocks will increase in size - they'll increase when they need to.  Does anyone actually take the position that it is better to let transactions back up than to marginally increase the incremental chain size?

If 20MB block size comes with pruning feature, i'm OK for this.  ;)
i want a counterparty ...  ;D

Finally it was about time!

BlockChain is  already almost 40 GB of size, but this change is inevitable. There is no problem with our current terabyte hard disks to storage it. And you dont have to run full node wallet like Bitcoin Core at all.
Electrum or MultiBit will do just fine.

Right now certain projects like Lighthouse decentralized crowdfunding are limited to 684 pledges (unusually low compared to kickstarter) and we can only process 4-7 transactions per second which isn't enough.

We are hitting the block limit more often these days and it is only going to get worse.

What is a serious problem already? The one who does not know what he/she is talking about is you ....

Blind following the blind going on!

Increasing the limit won't create bigger blocks on average, but allow for it to happen if it is needed. If we stay at 3-4 tps average the change will have no effect outside of occasionally allowing a 2MB block through because of a large fundraiser through lighthouse or a temporary spike in transactions.

There will likely be a middle ground found because most of the other core developers are wary of 20MB being needed. Perhaps if we give some other projects a bit of time to be released like the lightning network we won't need a 20MB limit.

Time to embrace Open Transactions, Ripple, Stellar or whatever for offchain transactions.

Edit:  I say go for it.  Go for the 20mb blocks.  I want to see what happens...  Just hedge your bets.

I don't think i understand your position.
Trading activity has always passed through offchain transactions..

So, do we really need ripple or stellar to make bitcoin work?
Why?

Nope, bitcoin works fine now , but lately we are starting to hit the blocksize limit more often.

This is what Bitcoin needs to make it awesome-

https://bitcointalk.org/index.php?topic=970822.0

Here's a Q&A with Gavin and Mike, at some point they discuss the bigger block issue if you're interested.

CoinScrum: QA with Gavin Andresen and Mike Hearn https://www.youtube.com/watch?v=RIafZXRDH7w

Agreed. This has been an issue for a while.

I don't know enough/haven't been around long enough to know one way better than the other, but I can say that I'm glad something is happening and all the back and forth can stop.  It works or it doesn't, time to find out.

http://gavinandresen.svbtle.com/why-increasing-the-max-block-size-is-urgent
http://gavinandresen.ninja/time-to-roll-out-bigger-blocks

Very good move :), Gavin is strong point of Bitcoin

one small step for future upscaling one Giant Leap towards complete centralisation  :-X

This should be interesting.

Correct me if I am wrong... Bill Gates said something like this ..." We WILL NOT need more than 64k of RAM "

He was clearly wrong... Look at where we are now. Satoshi made the same mistake, but the protocol allow for scalability, so it can be corrected.

If we go to mass adoption, we would have more transaction volumes and it's not currently possible with the current block size.

I would rather be pre-emptive than having a situation, where we sit with red faces, when we cannot handle huge transaction volumes and we get forced to do it in the future.  ???

You second link is the same one I started this thread with. :)

What mistake did Satoshi make?

and why satoshi didn't think about this? he predicted the increase in adoption for sure, because he predicted big farm, he should have created big blocks since the beginning, at least now with don't need any fork

You are wrong. He said nothing like that, and neither did Satoshi. Bitcoin originally had no block size limit; it was added later as an anti-DOS measure with the expectation that it would be increased if necessary.

Really? I never knew that! Who wrote the 1M limit into the codes? I have always assumed 1M was hard-coded in from the beginning, and miners/pools have been experimenting with the block size themselves to balance propagation speed and fees.

Bitcoin never had a block size limit originally ... it was added for DOS protection and satoshi himself said it would/could/should be upgraded in the future as needed .. (don't ask me for the post but I am sure someone else can pinpoint it)

Satoshi promised this 5 years ago

Thanks that is the one I was thinking off and there is another one where he discussed it IIRC..

As I understand it, Bitcoin blocks were always limited by MAX_SIZE (32 MiB).  It was in 2010 when MAX_BLOCK_SIZE (1MB) was introduced as a temporary anti-DOS measure.

Unfortunately, while many had high hopes that a robust mechanism could be devised to defeat the blocksize problem, no such proposals have survived close scrutiny.  Clever-sounding fee policy discussions (transactions becoming more expensive as blocks grow) were replaced with more general musings about votes (e.g. have full nodes report how long it takes them to download and verify blocks and adjust MAX_BLOCK_SIZE to target a sane duration).  Today, the blocksize limit remains in place and we seem to be left with few, very ugly options.

Disclaimer: I was quite new to Bitcoin at the time and am far from an authority on this topic.

Well you're right although I'm not sure if it was 32 MB. Someone older would have to verify if it had a 32MB limit at some point.
No the 1MB limit was not hard-coded from the begging. This is something that the majority seem to think, but it is wrong. We've reached that point in time where the measure, i.e. limit needs an increase.
We need to make enough room for future users and their transactions.

Even though getting a 1TB HDD is cheap today, what gets me worried is about the network.
Downloading a 40GB+ file isn't easy in the whole world(tons of places and countries have poor quality ISP with low speeds).
Light weight clients are nice and even I use them, but the blockchain is getting more and more centralized, what kills the initial concept o Bitcoin.

I think this has been discussed in-depth in the other thread and has been long in the planning stage. So I would say it's about time. Anyway it is something that we would need to prepare for the future because eventually we have to do it no matter how. If we want to go mainstream, increasing the size to absorb more transactions per block is the only way.

I don't that “kills” the initial concept of Bitcoin at all.

But no! The problem is not storage, but bandwidth. At least that's what many claim.

Bandwidth becomes a problem at over 1 MB every 10 minutes?  My moderate Internet connection can download 1 MB every 1/8th of a second.

Yes, but apparently there are people with 56k connection that want to run full nodes...

And now that I think about it, it's still enough to handle 20 MB every 10 minutes, isn't it?

Yeah I remember those 56k modem days. That's 56000 bits per second, not bytes. It wouldn't cope with 20MB blocks.

It isn't. It's only 2.5 MB (V.90) or 3 MB (V.92) every 10 minutes, assuming the node uploads as much as it downloads.

But didn't satoshi predict that this wouldn't be a problem? I remember reading emails of satoshi where he said it would never be a problem because technology would catch up as transaction volume grows.

The problem is, even a 1Mb needs 3 minutes to download one 20MB block.
Uploading a block with such connection takes more than 10 minutes, and downloading even only the headers will be frustrating.

Although today I have a 10Mb connection(its decent, could be better), I do worry about other people where internet connection is expensive.

Why worry?  Nobody is forced to run a full node.  The number of users who have bandwidth and storage to spare is nontrivial.  We are in no danger of a lack of users willing to run full nodes due to those concerns.

I think it's about time we change. we need to be future proof, and this is one big step towards that.

As for storage, the GB per $ price is quite cheap and is declining in the long run, for both mechanical and flash disks. I think the blockchain size is a bit of a non-issue. Also, the fact that the new limit is set at 20mb won't suddenly make the blockchain triple in size. The block sizes will remain the same. Our usage and the amount of transactions per block are the ones who will make the blockchain bigger, and I think more transactions is a good thing... :)

Reminder-

https://bitcointalk.org/index.php?topic=532.msg6306#msg6306


I don't predict that we will need to become this centralized but this was the initial plan.
One reason-Merkle tree Pruning scheduled for next bitcoin release(Core 0.11) allowing full nodes that are 1GB.

Satoshi said it would make sense to introduce pruning when the blockchain got too big. If a 20MB block size is going to be introduced it would be stupid to delay introducing pruning too.

Uh, what?  Substantiate your claim, provide proof.  There are over 6100 full nodes (https://getaddr.bitnodes.io/), and more added every day.  Making blocks bigger has no correlation with "centralization".

Bigger blocks ??? I think the block reward decrasing every year , can you please explain more what's coming

Kind Of Respect ,
Bitcoin Boy .

My home PC is one of those 6100 and I will have absolutely no problem to maintain it even the block size becomes 100MB

It's about having more data stored in the block. It has noting to do with the block reward

Block reward decreasing approx every 4 years...this proposed fork is intended to allow more transactions to be processed per second to help bitcoin grow and pay for the security.

For me it would be a nice feature, and it is very needed if bitcoin 'wants' to be mainstream. Because if we compare it with the credit card system we will see that 7 transactions per second - against - 2'000 transactions  per second (visa). In this way we will have a lot of transaction in one block and it will speed up the send of bitcoin.


I'm waiting to see this feature in the next releases and I am sure the majority will choose the right thing ;).

lol @ the sneaky little "kudos" buttons that mislead people into "liking" all his fallacies.


sounds like an Agent to me:

http://gavinandresen.ninja/it-must-be-done-but-is-not-a-panacea

Thanks for linking me to some great information from a very reasonable guy who has been critical to Bitcoins development. I will check it out.

Are you serious? Negative trust for saying your personal opinion http://techforum.it/styles/default/xenforo/smilies/asd.gif (amazing). When the feature will be released, everyone of us will choose if 'upgrade' his bitcoin core version or not (this is the good thing in a 'decentralization).

Let's play a game. Lets see who can spot all the Popescu sockpuppets that will inevitability pop up in this thread.
They'll turn up, oh yes, you can be sure of that. :D

ty for the link i didnt read it before;

well i think it will scaled slowly, yes, but i think i will need to start buying more hdds right now lol, just to use the btc core wallet :P

and yes i love to use the btc core wallet, so... xD

Some day in the future, you will need a data center to hold the whole blockchain, but that's still manageable, since by then each bitcoiner will own a data center  ;D

its not the increase in blocksize i was refering to but the way Gavin and co (particularly Mike) go about getting such done.

Yeah...  I don't know if some people are trolling, blind or clueless.

So we have a little less than one year to sell our Bitcoin before it becomes a clusterfuck. Ok, thanks for the info.

You don't want them? You can give them to me.

i think most rather dig their head in the sand where bitcoin is heading and how few control the code. and the interests of those few  :-X

https://twitter.com/petertoddbtc/status/595692609999347712

There s no other option then to reduce the size of the blockchain. BTW, when was the last time BTC chain forked or had any real issue? There are not kids playing in the sand.

Yup, and this is what a 250 megabyte hard drive looked like in 1979.

http://farm3.static.flickr.com/2743/4368314776_c8223ea75e_o.jpg

30 years later you can fit 250x250 megabytes on something the size of your finger nail.  When the blockchain needs a data center to hold it, a data center will fit on your desktop.

Just six posts later one has popped up...

LOL. Well spotted. :D

we are near that limit with silicon, we can't go much farther with shrinking, 10nm is the physical limit of the material, they need something else soon enough

Glass storage?
http://physicsworld.com/cws/article/news/2013/jul/17/5d-superman-memory-crystal-heralds-unlimited-lifetime-data-storage

last time i heard that they are working on graphene or nanotubes, as a next material, but glass could be another option(a storage in your glasses would be fabulous)

it's seems that they can push silicon to 7nm, but it will be too costly

They have been developing something called racetrack memory for years. However, it sounds like fusion, it's always another ten more years until they say they will get it working.

The capacity cliff

And why we can’t use fancy tricks to avoid it


https://medium.com/@octskyward/the-capacity-cliff-586d1bf7715e

Well that escalated quickly. What else to expect from a forum that has over 50% of shills?

One of the worst things is that Satoshi foresaw this. The longer we wait before a decision is made, the worse it gets.
The first transaction (with a fee) that doesn't get included in a block will cause panic.
Don't talk to me about the free market. Are we supposed to fight over who's going to be in the block? Should I pay a $10 fee to send $1? No, thank you. Any altcoin can do it for a much lower one.

With a 20MB block size limit you'll need a full TB per year if the network ever runs at full capacity. With a 1MB limit we're talking a more manageable 51GB/yr.

A 2TB HDD is selling at around US$70 currently. That's $0.096/day, or $2.92/month. And don't forget the cost will go nowhere but down.

Don't forget also that it will takes us a lot of years before each block is 20MB big.
I'd probably say 5+ years but that's just my guess. Memory and bandwidth will be much cheaper then than it is now.
Also it seems that nobody notices the pruning feature that we're going to have soon.