Please do not change MAX_BLOCK_SIZE

[gglon]

Set the limit to 10+MB and its almost certain that 5 years from now nobody will be able to run a bitcoin node at his home PC, through a DLS connection, not to mention via Tor.

I agree that decentralization is a priority. Though I think at least up to 100MB (if avg tx fee> $.1) we are perfectly safe in that matter, if we abandon Tor.

[behindtext]

at the rate transactions are picking up, it is only a matter of time before a full node needs to run on a small cluster of machines, e.g. a db machine, a tx verification machine, etc. imo, increasing the block size limit is a must.

unless countermeasures are taken, e.g. jgarzik suggestion of off-chain payments, the blockchain will continue to bloat, driving up tx fees. if the devs or miners choose to not raise the block size limit, users will complain and campaign for increasing the limit.

[kjj]

Set the limit to 10+MB and its almost certain that 5 years from now nobody will be able to run a bitcoin node at his home PC, through a DLS connection, not to mention via Tor.

I agree that decentralization is a priority. Though I think at least up to 100MB (if avg tx fee> $.1) we are perfectly safe in that matter, if we abandon Tor.

I would be very hesitant to abandon tor, and you should be too.

A lot of people are concerned that miners will soon find themselves in legal trouble.  I suspect their concerns are excessive, but what if they are right?  Right now, we can run the entire system over tor, which essentially makes it unkillable.  Until we are sure that we no longer need to be concerned about such things, or until a more efficient way to be unkilable comes along, we should not be making any changes which make truly anonymous operation harder.

[gglon]

I would be very hesitant to abandon tor, and you should be too.

I believe bitcoin may change the world. We can't be perfectly safe to achieve this.

A lot of people are concerned that miners will soon find themselves in legal trouble.  I suspect their concerns are excessive, but what if they are right?

Off-chain processors are even more likely to fall into trouble.

Right now, we can run the entire system over tor, which essentially makes it unkillable.

Unkillable - yes, but in case of "bitcoin war" it will be suitable only for drugs and criminal activities. So essentially it will lose >90% value anyway.

On the other hand, if it gains people acceptance,  not only will it be inconvenient for gov's to shut it down, but also it may gain much value.

Until we are sure that we no longer need to be concerned about such things, or until a more efficient way to be unkilable comes along, we should not be making any changes which make truly anonymous operation harder.

We will never be sure, and people acceptance would be pretty much efficient.

[justusranvier]

Virtually every bitcoin user already uses third parties, and trusts them to some extent.  Doing so is a trade, cost for risk.  In the future, I expect it to be even more complicated.  Imagine a credit card denominated in bitcoin.  With that, you'd be paying for them to assume transaction risk for you, while right now the model is mostly the opposite with you getting a discount for assuming the risk that the third party may bail with your bitcoins.

Right now trusting third parties is an optional choice that a user can make. In a scenario in which the demand for transactions exceeds the protocol-specified limit they won't have a choice - the only way average people will be able to use Bitcoin at all is to let a third party hold their coins for them. At this point Bitcoin is dead because it's become just another Dwolla or PayPal.

The mad scientist in me thinks that we should wait until the 1 MB limit is met and sustained for a while before we raise it.  That way we'd get some real world experience with how the system (including the people!) react.

Imagine an accelerating car driving into a brick wall. One day the number of users and transactions will be increasing exponentially and then suddenly the growing user base will be rationed to a fixed number of daily transactions. It won't matter how much the users are willing to pay in transaction fees or how much the miners are willing to provide a higher transaction rate - the network will be limited to a constant number of transactions. It's madness to blindly assume that Bitcoin will continue to be useful and in adoption after its use case is fundamentally reversed.

Right now the demand for transactions is below MAX_BLOCK_SIZE, so effectively we have no limit. What we're looking at right now is how users and miners behave with "infinite" block sizes. Users are willing to pay fees, and the fee revenue grows proportionally with the transaction rate. We don't have to speculate about this because the real data is available. Allowing Bitcoin growth to suddenly hit an artificial limit really is playing mad scientist with what will be at that point a multi-billion dollar economy. Incidentally that kind of central planning is exactly what many people are adopting Bitcoin to getaway from and you're ready to pull the rug out from under them right when they least expect it.

[amincd]

As I have read it many times before, and I completely agree with it: Bitcoin is not designed for micro-transactions.

With a 1 MB block size limit, Bitcoin will not be usable for normal transactions as a global currency. It would only allow 7 txs per second, meaning only huge value transactions (e.g. >$5,000) would be economical.

The network does not scale, we have a worldwide economic crisis and the Moore's law does not seem to be any longer applicable; our internet connections got stuck at what a DLS's copper can do, and the CPU's also don't seem to be getting the speed as much, as they were 10 years ago.

Moore's law is continuing, with the number of transistors on integrated circuits increasing exponentially. Internet connections have gotten much faster since 10 years ago too.

So why not to just keep this point at the 1MB, as Satoshi originally designed?

Satoshi originally designed Bitcoin with the vision that full nodes would eventually handle blocks that are hundreds of MBs in size and be run by specialists:

http://www.mail-archive.com/cryptography@metzdowd.com/msg09964.html

At first, most users would run network nodes, but as the network grows beyond a certain point, it would be left more and more to specialists with server farms of specialized hardware.  A server farm would only need to have one node on the network and the rest of the LAN connects with that one node.

The bandwidth might not be as prohibitive as you think.  A typical transaction would be about 400 bytes (ECC is nicely compact).  Each transaction has to be broadcast twice, so lets say 1KB per transaction.  Visa processed 37 billion transactions in FY2008, or an average of 100 million transactions per day.  That many transactions would take 100GB of bandwidth, or the size of 12 DVD or 2 HD quality movies, or about $18 worth of bandwidth at current prices.

If the network were to get that big, it would take several years, and by then, sending 2 HD movies over the Internet would probably not seem like a big deal.  

Originally there was no 1 MB limit in the Bitcoin code. The 1 MB limit was a temporary measure put in place until a better way to control bloat could be found. It was never intended to be permanent.

If you don't increase MAX_BLOCK_SIZE people will naturally start using BTC payment processors, which will take the load off the net, and which has to eventually happen anyway.

People will have to use BTC-banks, not payment processors. We're not talking about Bitpay simply converting your BTC to dollars for some merchant. We're talking about having to store your BTC at an e-wallet, and send BTC-credit, backed by that e-wallet, whenever you want to transfer 'BTC'. This would be a credit based economy that would have a tendency toward centralization.

There's a long discussion on the block size limit here that I recommend you read to see counter-arguments to keeping blocks limited to 1 MB:

https://bitcointalk.org/index.php?topic=208200.0

Speaking of scale, we know that at some level Bitcoin just can't be decentralized and still have every transaction on the blockchain - you just have to read the wiki page on Scalability to see how VISA-level volumes require a system that's hardly decentralized, let alone resistant to government control.

Only if you define "decentralized" as home PCs being able to run full nodes, which is not the way Nakamoto defined it. Having people reliant on handling BTC-promissory-notes backed by centralized BTC-banks is giving up on BTC-decentralization altogether, as well as the original mission for Bitcoin of being a "digital cash".

Here's an interesting thing I just conjured up:  From a doc 'UnitedStatesComp.pdf'

Fedwire processed an average of nearly 430,000 payments per day in 2000. The total value of
transfers originated during 2000 was USD 380 trillion. The distribution of the value of these payments
is not uniform. The median Fedwire payment during 2000 was approximately USD 25,000, and the
average payment was approximately USD 3.5 million.

To save one the math, I get about 5 tps.

Seems to me truly a no-brainer to consider various forms of off-chain transactions vs. trying to shove everything onto the blockchain.  Even with 'ultra-prune'.  The only real question is when.  A median of $25k and average of $3.5M is pretty alarming...though we are talking about $380T here...  It cannot be said that I am not a 'USD user' by virtue of the fact that I've never performed a Fedwire transaction.  The same would stand for Bitcoin.

You have handled cash, which makes you a 'USD user'. Also, while you haven't use Fedwire, you have used banks regulated by the federal government and a payment network that relies on a centralized clearing house run by the Federal Reserve: the ACH. Unless we have a similar government-regulated and insured BTC-banking system, and a centralized BTC clearing house, BTC-credit held at various banks won't be equivalent to BTC.

[piotr_n]

You cannot argue with one thing from it; by increasing the blocks size, just to keep up with the growing number of txs, you will eventually end up with a system where a bitcoin node can by run only by rich people and only in data centers.
Set the limit to 10+MB and its almost certain that 5 years from now nobody will be able to run a bitcoin node at his home PC, through a DLS connection, not to mention via Tor.

thats not true. 10mb can likely be sustained virtually forever. both bandwidth and storage space grow fast enough to handle this.

Really?  Then where live must be far behind the rest of the world, since the internet speed does not seem to have improved here over the last couple of years. Thus no reason to assume that it will improve much in the next couple of years...

As for the CPU usage, I think 7 tx/sec seems to be a tolerable maximum, since it adds up to about 15 elliptic curve verify ops, per second.
If you have a CPU with 2 cores, and about 2 GHz clock - it can stand it in a background.
But if you make it 150 EC ops/s, leaving the came CPU, you are just killing all these poor people's home PCs - and that's not even counting the bandwidth, nor the memory requirements.

So yeah, if you want to make a bitcoin node available only for the rich - then increasing the block size is probably the best way to go
But if your goal is to just handle more transactions per second, you should rather try to push the inevitable centralization into more democratic directions - ones that don't give advantages to the rich. Meaning: many centralized payment processors who settle their balances with each other through the bitcoin blockchain - this seems to be a natural and democratic way to go forward, IMHO.

[DeathAndTaxes]

at the rate transactions are picking up, it is only a matter of time before a full node needs to run on a small cluster of machines, e.g. a db machine, a tx verification machine, etc. imo, increasing the block size limit is a must.

Not really.  CPU is not much of a bottleneck.  RAM requirements may eventually require machines with above average memory but even they are modest through 70tps or so.   Storage is dirt cheap and getting cheaper all the time so it is less of a concern for existing nodes.  Now bootstrapping a 400GB blockchain for new nodes will require some optimization but it isn't impossible.  Bandwidth is the key limiter.  Residential connections only have so much throughput and while improved (optimal) tx and block relay procedures can reduce that average nodes bandwidth requirements by 75% or more pretty soon one is going to hit the limit of what a residential connection can handle.  Assumming sufficient disk space and bandwidth I wouldn't see the need for clusters of machines until tps reaches well into the hundreds of thousands per second (which is likely never).

[gglon]

As for the CPU usage, I think 7 tx/sec seems to be a tolerable maximum (...)

see: https://en.bitcoin.it/wiki/Scalability#CPU

So yeah, if you want to make a bitcoin node available only for the rich - then increasing the block size is probably the best way to go

But if your goal is for it to just handle more transactions per second, you should rather try to push the inevitable centralization into more democratic directions - ones that don't give advantages to the rich. Many centralized payment processors who settle their balances with each other through the bitcoin blockchain - this seems to be a natural and democratic way to go forward, IMHO.

Making bitcoin full node available for rich enough to sustain 100MB is enough in terms of decentralization. And there is no advantage to the rich apart from the fact, that normal users need to use servers for tx handling - still much safer than using offchain processors. So this is democratic direction.

[piotr_n]

As for the CPU usage, I think 7 tx/sec seems to be a tolerable maximum (...)

see: https://en.bitcoin.it/wiki/Scalability#CPU

"4000 tps is easily achievable" - of course... wiki always knows the truth
And then one can only wonder why a 250KB block needs over one second to get verified by my i5 CPU, occupying 100% of it, while doing it..

Making bitcoin full node available for rich enough to sustain 100MB is enough in terms of decentralization. And there is no advantage to the rich apart from the fact, that normal users need to use servers for tx handling - still much safer than using offchain processors. So this is democratic direction.

I think you are not aware of what kind of world we are living in.
If only rich and powerful can run a bitcoin node - then it will become very easy for a government to shut it down. Especially after you disable support for Tor.

[gglon]

If only rich and powerful can run a bitcoin node - then it will become very easy for a government to shut it down. Especially after you disable support for Tor.

Min requirements for full node at 100MB is ~1mb dsl connection, perhaps 10x more for miners, and 1hdd/year.  You don't need to be especially rich to meet this.

[justusranvier]

If only rich and powerful can run a bitcoin node - then it will become very easy for a government to shut it down.

You're setting up a false choice by presenting a scenario in which only the rich and powerful being able to run a Bitcoin node, and then you propose a solution (limited block size) which results in only the rich and powerful being able to transact on the blockchain and call this a solution.

Then there's the making up facts in order to support your position.

So why not to just keep this point at the 1MB, as Satoshi originally designed?

That Satoshi originally designed a 1 MB limit is wholly false, and this has been pointed out to you, and you have yet to acknowledge it. The odds that you're arguing in good faith is low at this point.

People who are genuinely interested in solving the problem of full node resource usage, instead of just using it as a red herring to further some ulterior motive, can contribute bitcoins and/or code towards actually fixing the problem.

[piotr_n]

If only rich and powerful can run a bitcoin node - then it will become very easy for a government to shut it down.

You're setting up a false choice by presenting a scenario in which only the rich and powerful being able to run a Bitcoin node, and then you propose a solution (limited block size) which results in only the rich and powerful being able to transact on the blockchain and call this a solution.

Then there's the making up facts in order to support your position.

So why not to just keep this point at the 1MB, as Satoshi originally designed?

That Satoshi originally designed a 1 MB limit is wholly false, and this has been pointed out to you, and you have yet to acknowledge it. The odds that you're arguing in good faith is low at this point.

People who are genuinely interested in solving the problem of full node resource usage, instead of just using it as a red herring to further some ulterior motive, can contribute bitcoins and/or code towards actually fixing the problem.

I don't like the way you're talking to me, but out of the respect for other people reading this thread, I'm not going to reply on your void accusations.
But you are wrong - all the way.

[Peter Todd]

People who are genuinely interested in solving the problem of full node resource usage, instead of just using it as a red herring to further some ulterior motive, can contribute bitcoins and/or code towards actually fixing the problem.

I don't like the way you're talking to me, but out of the respect for other people reading this thread, I'm not going to reply on your void accusations.
But you are wrong - all the way.

FWIW piotr_n UTXO indexes don't help scalability with regards to bandwidth because you still have to receive every transaction on the P2P network to mine profitably or validate properly.

Although if we're going to go off accusing people of ulterior motives, anyone want to figure out what, if any, services/companies justusranvier runs? It's interesting how amincd runs btctip for instance.

[justusranvier]

Although if we're going to go off accusing people of ulterior motives, anyone want to figure out what, if any, services/companies justusranvier runs? It's interesting how amincd runs btctip for instance.

All you have to do is ask. I'm working on an educational project to introduce Bitcoin to non-computer geeks. It hasn't launched yet though.

The difference between a mistake and a lie is that people correct factual mistakes when they are pointed out. Continuing to say that the 1 MB block limit was part of Bitcoin's original design is a lie.

I only start to publicly speculate about ulterior motives when someone is deliberately spreading falsehoods.

[piotr_n]

The difference between a mistake and a lie is that people correct factual mistakes when they are pointed out. Continuing to say that the 1 MB block limit was part of Bitcoin's original design is a lie.

It's not a lie, man.
1 MB block limit - it's there and has been there for years, whether you like it or not.
Was it Satoshi himself who added it - who cares? Who is Satoshi, anyway?
The only thing that matters is that some people don't want it to grow bigger, and they have an actual valid reasons. Very important reasons.
And you accusing me of "making up facts" - it's just silly

[tvbcof]

...
Though there is a potential market for something like torcoin (paracoin), it's not what >90% bitcoin users (miners) would want to have; at least under present circumstances.
...

For the record, Tor was not really on my mind for this.  I've never convinced myself that Tor is a sufficiently reliable framework in 'war time' environments.

Paracoin is mostly envisioned as a platform upon which to reasonably experiment with constructs which would push core infrastructure to the network's edge...and beyond if need be.  These might include rewards proportional to desired diversity in terms of geo and political situation, transaction messaging protocol, peer count, and such.

My latest thoughts have to do with 'two-pass' mining where the 'tip-blocks' are secured with suite of diverse and non-predictable algorithms while the trailing blocks are locked with high difficulty sha256.  Hopefully this could reward both those running FPGA (hopefully at home) while anyone with an investment in current ASIC and what-not could continue to participate and secure the solution.

But anyway, I would strongly prefer that Bitcoin proper remain the 'gold standard' as a source of truth in distributed crypto-curency land.  'paracoin' and things like it are only a fall-back should Bitcoin go the way of PayPal due to excessive commercialization and associated capture.  It probably understates things to sat '>90%'.  It's probably closer to '>99%'.

  Edit:  More clear about 'edge'.

[justusranvier]

The difference between a mistake and a lie is that people correct factual mistakes when they are pointed out. Continuing to say that the 1 MB block limit was part of Bitcoin's original design is a lie.

It's not a lie, man.
1 MB block limit - is there, has been there for years, whether you like it, or not.
Was it Satoshi himself who added it - who cares?

Obviously you care, or at least you think the people you are trying to convince care, because you relied on as the basis of your appeal to authority.

So why not to just keep this point at the 1MB, as Satoshi originally designed?

So it's important when it supports your point, but then we shouldn't worry about it if it turns out to be false?

[Peter Todd]

I've said many times before that Satoshi was a man, not a god.

Bitcoin, as originally released, had a 32MiB blocksize limit and it was implemented in a way that makes me wonder if Satoshi didn't realize what an issue the blocksize limit would be. He later changed that limit to 1MB, and did so relatively secretly.

In any case what matters is how we can make Bitcoin scale now while remaining decentralized and censorship resistant, not what Satoshi thought Bitcoin should be four years ago.

[piotr_n]

In any case what matters is how we can make Bitcoin scale now while remaining decentralized and censorship resistant, not what Satoshi thought Bitcoin should be four years ago.

Exactly.
So my idea is: leave the limit as is, and let the fee market to work. This will naturally create incentives for payment processors to appear and compete with each other. Just like mining pools do ATM. It even makes a lot of sense that the mining pools would become payment processors.