Bitcoin Forum
March 28, 2024, 06:52:00 PM *
News: Latest Bitcoin Core release: 26.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 2 3 [4]  All
  Print  
Author Topic: Why does there need to be a limit on amount of transactions?  (Read 3735 times)
DeathAndTaxes
Donator
Legendary
*
Offline Offline

Activity: 1218
Merit: 1063


Gerald Davis


View Profile
May 31, 2013, 12:47:05 AM
 #61

Yes there are ways to optimize traffic however they aren't implemented yet so TODAY those limits hold.  Also transmitting a hash (64 bytes) vs transmitting a tx (~400 bytes on average) while smaller isn't an order of magnitude smaller.  So take whatever realistic limit exists based on available resources and even hyper optimized the same limit still exists at 5x that transaction limit.

Don't get hung up on the details.  The point is as tx volume increases bandwidth is the tightest bottleneck.  As nodes can't or won't handle that the number of full nodes will decline.  Higher transaction volume when it breaches what is "reasonable" (and yes there is some gray area on what is reasonable) to the average node will result in a centralization as a result of the costs to run a full node.  Optimizations help but you can't get blood from a stone.   Nobody is going to be able to run a full node on the AVERAGE residential connection with a transaction speed of say 5,000 tps.


Quote
Just take a look at the amount of competitors that show up in places where banking regulations are less burdensome, like Panama, and compare it with other places (relatively to the country's population and GDP sizes)
There are far more bitcoin nodes today than banks in Panama.  Regulation or not when the burden on a node rises there will be less nodes.  That is a form of centralization.  An open unlimited network "may" reach a good compromise or it may not and this is a billion dollar project.  Their are no "oops" guess it didn't lets hit the reset button.  Given that expanding Bitcoin is somewhat like conducting maintenance on an aircraft in flight it might be worthwhile to move cautiously.   
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
"Governments are good at cutting off the heads of a centrally controlled networks like Napster, but pure P2P networks like Gnutella and Tor seem to be holding their own." -- Satoshi
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
1711651920
Hero Member
*
Offline Offline

Posts: 1711651920

View Profile Personal Message (Offline)

Ignore
1711651920
Reply with quote  #2

1711651920
Report to moderator
d'aniel
Sr. Member
****
Offline Offline

Activity: 461
Merit: 251


View Profile
May 31, 2013, 02:31:37 AM
 #62

Yes there are ways to optimize traffic however they aren't implemented yet so TODAY those limits hold.  Also transmitting a hash (64 bytes) vs transmitting a tx (~400 bytes on average) while smaller isn't an order of magnitude smaller.  So take whatever realistic limit exists based on available resources and even hyper optimized the same limit still exists at 5x that transaction limit.
Transaction hashes are 32 bytes, but even only the first few bytes are enough to identify the tx in the mempool.  So it's actually about two orders of magnitude less data than sending full txs.
n8rwJeTt8TrrLKPa55eU
Hero Member
*****
Offline Offline

Activity: 588
Merit: 500



View Profile
May 31, 2013, 03:30:12 AM
 #63

Given that expanding Bitcoin is somewhat like conducting maintenance on an aircraft in flight it might be worthwhile to move cautiously.    

This.  Anyone who has worked in the aerospace/airline industry on mission critical systems knows that you don't make changes on gut feelings and optimism.  When people's lives are at stake, you are extremely conservative, assume the worst, triple check, and try to get 100% unanimity on any risky assessment.

The "hard 1MB" position perhaps has some adherents not necessarily out of irrational conservatism or ideological extremism, but to counterbalance people who appear to seriously want to dump the limit altogether in one go, and hope things will just work out, with no empirical evidence beyond thought experiments and forum debates.

In spite of the usual disclaimers, Bitcoin is no longer toy money.  Its stored value is now larger than the GDP of many countries.  Lots of simulation, hard data, and stress testing will be needed before prudent people will go along with changes that risk unintended consequences on a $1B+ economy.

edmundedgar
Sr. Member
****
Offline Offline

Activity: 352
Merit: 250


https://www.realitykeys.com


View Profile WWW
May 31, 2013, 06:21:09 AM
 #64

Given that expanding Bitcoin is somewhat like conducting maintenance on an aircraft in flight it might be worthwhile to move cautiously.    

This.  Anyone who has worked in the aerospace/airline industry on mission critical systems knows that you don't make changes on gut feelings and optimism.  When people's lives are at stake, you are extremely conservative, assume the worst, triple check, and try to get 100% unanimity on any risky assessment.

The "hard 1MB" position perhaps has some adherents not necessarily out of irrational conservatism or ideological extremism, but to counterbalance people who appear to seriously want to dump the limit altogether in one go, and hope things will just work out, with no empirical evidence beyond thought experiments and forum debates.

In spite of the usual disclaimers, Bitcoin is no longer toy money.  Its stored value is now larger than the GDP of many countries.  Lots of simulation, hard data, and stress testing will be needed before prudent people will go along with changes that risk unintended consequences on a $1B+ economy.



I think most people would agree with that, but it's worth saying here that keeping the ceiling where it is as we get ever closer to it isn't necessarily the cautious move.

To stick with the aircraft analogy, we're flying at 1000 meters towards a 1500-meter-high mountain. If we do nothing, we hit the mountain. There may be some legitimate concerns about whether the plane will be OK at higher altitude, but there are also serious concerns about what happens when the plane hits the mountain. The fact that we've been flying along happily at this altitude over open sea for quite a while before we actually got to the mountain doesn't mean that everything will be OK at the same altitude when we do hit it.

Specifically, we've always had reasonably low transaction fees. We have no idea what would happen to the economy with high transaction fees, and there are some very plausible worst-cases, like Bitcoin losing in the marketplace to an alt-coin or another technology, and most of that $1B+ going up in smoke. The people advocating keeping the limit low are generally optimistic on how great it might be to hit it (we'll land softly in a blossoming forest of off-chain solutions!), and also of the opinion that we're going to hit a mountain sooner or later anyway so it may as well be sooner. They may even turn out to be right. But they're bold, not cautious. The cautious move is not to hit the mountain.
justusranvier
Legendary
*
Offline Offline

Activity: 1400
Merit: 1006



View Profile
May 31, 2013, 06:24:30 AM
 #65

The people advocating keeping the limit low are generally optimistic on how great it might be to hit it (let the off-chain solutions bloom!), and also the opinion that we're going to hit a mountain sooner or later anyway so it may as well be sooner. They may even turn out to be right. But they're bold, not cautious. The cautious move is not to hit the mountain.
They also tend to drastically underestimate exponential adoption rates. We don't have a few years before we hit the mountain. The right time to start addressing this was last year.
caveden
Legendary
*
Offline Offline

Activity: 1106
Merit: 1004



View Profile
May 31, 2013, 06:39:22 AM
 #66

but to counterbalance people who appear to seriously want to dump the limit altogether in one go, and hope things will just work out, with no empirical evidence beyond thought experiments and forum debates.

Sigh...
It's not a "hope", it's an aprioristic certainty. You don't need central planning to avoid "nefarious market cartelization", just study economics if you don't believe me.

Talking Bitcoin specifics: it's easy to spot an attempt of spamming by another miner. Its blocks will contain a large percentage of unknown transactions. So, just create soft limits to censor blocks with many unknown transactions. Say, if a block contains more than 10% of unknown transactions, don't build on top of it unless it's already 1 block deep. If it contains >20%, wait for it to be 2 blocks deep etc. Obviously the percentages and depths should be configurable.
You can also add such limits on the block size itself. Larger than X, wait for Y depth at least. Multiple (X,Y) configurable pairs.
Oh, and as a bonus, such soft limits would create an incentive for miners to relay and broadcast transactions they receive - today they have no incentive other than the counter-incentive of having to patch the software. If they keep transactions they receive from SPV nodes for themselves, they might get their block orphaned.

It's quite visible that miners would only slightly raise their limits when they believe the gains from adding more paying transactions would outcome the potential losses from orphanage. That's spontaneous regulation, transaction space supply adapting to its demand.
And it's absurd to claim that a remote, unclear chance of kicking out very low bandwidth miners would be so attractive as to make large bandwidth miners take the risk of losing their entire block on orphanage.

Please, people, being p2p is the greatest feature of Bitcoin. P2P is all about spontaneous order - an actual verifiable fact, not a mere "hope". How can you claim to support the first and largest p2p currency and yet be against spontaneous order?

Nagato
Full Member
***
Offline Offline

Activity: 150
Merit: 100



View Profile WWW
May 31, 2013, 09:40:06 AM
 #67

Except that the block size issue is not akin to free market dynamics.

In a free market, when someone makes a purchase, the only parties directly affected are the purchaser and the seller.

In the blockchain world, when someone publishes a 1GB block, the price is paid by every full node while the reward is only
collected by the publisher. This dynamic is closer to private profit, public loss.

There is much incentive for a well connected miner to publish a large block for 3 reasons
i) He and only he gets more txn fees. The larger the block, the more revenue he gets.
ii) Price for large block is paid by everyone (read: my competition today and tomorrow)
iii) In the event the huge block is not orphaned, he gets a headstart to mine the next block while the rest of the network sits idle trying to download his block. This leads to him having a higher probability than his actual hashing power to find the next block(go to i).

Now lets look at empirical evidence.

P2Pool has an insignificant share of hashing power even though miners get to keep 100% of all earnings vs mining pools which take a cut or txn fees.

Why?
Because the cost of running a full node outweighs the the revenue loss from mining with a pool.
And this is with the average block being <250KB.

There is no question that mining with a pool brings down mining cost because you don't need a fat pipe and a decent PC with a large HD/RAM to run bitcoind.
Decentralised systems have redundancies which leads to much higher costs when compared with centralised systems and the free market may eventually force centralisation in Bitcoin as well. The question is whether we want Bitcoin to centralise towards off-chain solutions for smaller transactions(low blocksize with many miners) or centralise the currency itself to allow on-chain txns for everyone(big blocksize with few large miners).

Personally i think keeping the Bitcoin protocol decentralised to be much more important than keeping its direct transactional capabilities decentralised. However i understand that there is a risk that people will just start using off-chain solutions as money just as paper notes were used instead of gold.

Ideally, the community takes the middle ground and increases the block size slowly to keep pace with bandwidth increases. There is probably an inflection point for bandwidth where blocks will be large enough for the global population, but we are not there yet.

caveden
Legendary
*
Offline Offline

Activity: 1106
Merit: 1004



View Profile
May 31, 2013, 12:24:33 PM
 #68

Except that the block size issue is not akin to free market dynamics.

In a free market, when someone makes a purchase, the only parties directly affected are the purchaser and the seller.

In the blockchain world, when someone publishes a 1GB block, the price is paid by every full node while the reward is only
collected by the publisher. This dynamic is closer to private profit, public loss.

Dude, this a repeated scenario in economic theory. The "fear" that free markets cannot internalize costs when there's the possibility of a tragedy of the commons. It's similar to security in the meatspace.
And guess what? It's perfectly possible to eliminate the tragedy of the commons risk through spontaneous order, as long as property rights are established and respected.

And btw, private profit and public loss only happens when the state comes into the scene.

There is much incentive for a well connected miner to publish a large block for 3 reasons
i) He and only he gets more txn fees. The larger the block, the more revenue he gets.

Nagato, if there are real transactions paying large fees to get included, this represents real demand. Miners better attend it or Bitcoin jams!

And as I explained twice on this thread already, the risk of hitting some soft limits would make miners be prudent on this. They would only increase their blocks when there's enough consensus, or when the demand is so strong that it's worth the risks. In both cases we are fine.

P2Pool has an insignificant share of hashing power even though miners get to keep 100% of all earnings vs mining pools which take a cut or txn fees.

Why?
Because the cost of running a full node outweighs the the revenue loss from mining with a pool.

Please, I and many others run a full node without getting nothing in return.

AFAIK P2Pool is not very popular because it allegedly has large stale rates. I don't know if this claim is factual.

Personally i think keeping the Bitcoin protocol decentralised to be much more important than keeping its direct transactional capabilities decentralised.

Both things will always be the case, if you remove the hardcoded constant limit.

Ideally, the community takes the middle ground and increases the block size slowly to keep pace with bandwidth increases.

But that's precisely what I'm saying! Block size should be controlled by everybody, with their choices and plannings, not by a centrally imposed formula.
cdog
Hero Member
*****
Offline Offline

Activity: 1036
Merit: 500


View Profile
May 31, 2013, 08:21:15 PM
 #69

Personally I think the solution is more than one coin.

Bitcoins for storing value long term and bigger payments.

Litecoins for day to day transactions and purchases under $100. 
edmundedgar
Sr. Member
****
Offline Offline

Activity: 352
Merit: 250


https://www.realitykeys.com


View Profile WWW
May 31, 2013, 11:16:47 PM
 #70

Personally I think the solution is more than one coin.

Bitcoins for storing value long term and bigger payments.

Litecoins for day to day transactions and purchases under $100.  

If nearly all the actual commerce was happening in Litecoins, and Litecoins had solved the scaling problems that Bitcoin hadn't, it's not really clear that people would still want to hold Bitcoins.

The normal thing to do when you run into serious scaling problems is to shard. Sharding is always a big PITA, so you don't do it until you absolutely have to. (Half the payment has confirmed, but the other half hasn't yet, because you're spending coins from two different shards...)

But right now Bitcoin is scaling fine. It's not radically decentralized, and hasn't been for a long time, but that ship sailed when people started mining on expensive, dedicated hardware rather than using the spare CPU cycles on boxes they were running anyway. Capital-intensive commodity businesses have huge economies of scale. They are never cottage industries. You can't run a profitable steelworks in your garage. We may not like these economic facts, but tinkering with irrelevant things like the block size isn't going to change them.
Pages: « 1 2 3 [4]  All
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!