Topic: Block Size/Transaction Speed/Mainstream Adoption (Read 1521 times)

Thanks for the posts, I think I am starting to see/understand all the fuss about XT, I will get to that in a minute.

I agree, No Well Founded Technical reason not to reduce the blocktime/interval to 5 minutes.

So, really the main issue about reducing the blocktime/interval is that it will create more orphan blocks and the time it takes to recover all the network to mine on the correct block.  So, if we reduce the blocktime/interval we will have more orphan blocks.  For example if we use 5 min, the difficulty would have to be adjusted to make 5 min blocks, no real issue.  Using 5 min blocks would create roughly twice as many orphan blocks - but with the mining reward being halved to 12.5 btc (currently at 25), the btc loss to miners for orphan blocks is the same.  The real issue then is recovery time to get everyone back on the correct block.  I am not sure how to evaluate that issue.  From blockchain.info https://blockchain.info/charts/n-orphaned-blocks?showDataPoints=false×pan=&show_header=true&daysAverageString=7&scale=0&address= It looks like the 7 day average of orphaned blocks varies between 0.5 to 3 orphan blocks per day.

Information about time it takes a new block to be propagated throughout the network is in seconds.  I had assumed that this would be in milliseconds or a second or two, but currently it is much larger.  It says the mean (~average) time it takes a node to see a new block is 12.6 seconds and by 40 seconds 95% of the nodes have the new block.

Also, INCREASING THE BLOCK SIZE, for every kb increase in size adds 80ms delay until the majority of nodes know about the block.  So going from 1mb to 8mb is 7168kb increase would increase average block propagation time from 40 seconds to 6 minutes, 13 seconds.  (now that is talking about a block that is 20kb in size and 1 kb increase in size adds 80 ms, I am wondering if this 1 kb increase to a 1 mb block would cause greater than 80 ms propagation delay?)
Those stats are from
http://www.cs.ucsb.edu/~rich/class/cs290-cloud/papers/bitcoin-delay.pdf

QUESTION: Does it look like like Gavin is trying to centralize the bitcoin network?

There were some posts on reddit over the weekend about an old conversation (2011?) Gavin saying to Satoshi, that once bitcoin grows and they increase the blocksize it will upset people running nodes on their home PC.

OK, just speculating here:
So, if one with power wanted to centralize bitcoin to keep control of it, I am not sure bitcoin would really survive that, but if they knew a way they could keep it going and gain control it, It seems one of the best ways to start to accomplish this would be to make the blocks very large so that you could centralize a small network of nodes and miners to put controls you want in place.

Tin foil hat, or not?

Block propagation time is the sum of two components, a fixed transit delay (ultimately bounded by the "speed of light") and a size related component (roughly equal to the block size divided by the speed of the path bottleneck).  One of the reasons for having a large average interval between blocks is to ensure that the fixed transit delay does not play a major role in the profitability of a mining operation.  Absent this factor, we would have the situation as it exists in New York City where one needs to be located within milliseconds (or less) to do profitable high frequency trading.

With a small average interval between blocks the fixed transit delay would begin to dominate the probability of winning an orphan race.  This would provide an economic edge for geographic concentration of bitcoin.  This would not only result in the geographical concentration of mining pools in a small area of the world, it would also drive geographical concentration of hash power as well into an area where there was ample cheap electricity.

My opinion is that there is no well-founded technical reason not to reduce the average block interval instead of increasing the block size, if you accept the later.

Every change to the Bitcoin protocol benefits some parties more than others. If you want to improve Bitcoin core constants, you will generate an imbalance.

But reducing the block interval actually benefits all the community so much that any imbalance it creates becomes irrelevant.

I've been pushing for block rate reduction to at least 5 minutes since long ago.

With some technical changes, such as implementing the DECOR+ protocol, you could increase the block rate and also even reduce the orphan rate, since you can prevent selfish mining.

"Throughput" and "Response time" are two different concepts.  Increasing the maximum blocksize today may have little or no effect on the time to get a transaction confirmed today, because the network is not operating near its peak throughput.  However, if next week the load should just happen to be much higher, then queueing will take place and confirmation may take hours or days.

I understand and do some of the things you describe, gyft cards etc. but most ppl wont - its a hassle with no benefit to mainstream.

If we have to covert our bitcoins to fiat to buy anything, why have bitcoin?

If no one uses bitcoin the bitcoin value = $0.  I assume most all of us here would like to see btc> $1000.

I do inderstand the issue of more orphans being created with a 24 sec block, but what about a 3 min block?  Or somewhere less than 10 min.

I do understand, as blocktime approaches zero, having 100 ph may not be any better than 10 TH.  But someone with 10 TH would have the same odds of finding a block on a 400ph network regardless if blocktime is 5 or 10 minutes.

There is an issue of slow confirmation times that just increasing block size will never make this faster - it will only prevent it from slowing down.

So my 24 sec block was an extreme, what would be wrong with a 5 minute block?  I see it would probably double amount of orphan blocks, but what else?

I am sure miners would be ok with double orphan blocks if btc >$750?  but no guarantee of that.

Imagine purchasing electronic gift cards days in advance (waiting for as many confirmations as needed) and then simply using those for your instant purchases.

Imagine paying for things with credit cards (or other forms of credit) and then paying off the monthly balance with bitcoins.

Imagine a merchant asking for some identification along with a no confirmation purchase.

Imagine merchants simply not requiring confirmations for anything less than $100. I'd imagine other kinds of fraud are much easier to accomplish than double spending bitcoins and merchants already deal with those kinds just fine.

There are plenty of ways to bypass the need for confirmations, it just takes a bit of imagination.


Is mainstream use of Bitcoin a goal of yours? Why?


Do you need the benefits of decentralization when paying for your dish subscription? Probably not. Every transaction under the sun doesn't need the censorship-proof capabilities of the block chain. All we need is for those capabilities to exist for the transactions which might be censored. The very fact that a censorship-proof payment option exists will probably cut back on censorship itself (as it's futile).

Also, a shorter block time will mean that the network will suffer from latency (nodes and miners would suffer even more from bandwidth and internet issues).
If you think that increasing the block size is difficult, imagine changing the block time. I guess people need to realize that bitcoin cannot be used for everyday transactions, not right now. It needs more development or implementations built on top of it to be able to handle such traffic.

Well shorter block times means that the network is more likely to produce orphans and fork, creating longer and harder blockchain reorgs. Since it would be easier to pull off a double spend (need less hash power to produce a block), the amount of confirmations required to guarantee that a double spend had not occurred would increase, so you may still end up waiting 20 minutes for X number of confirmations to guarantee that a double spend did not occur.

Yep, I understand.  even if your tx gets in the block being mined when you send the tx, you still have 20 min wait to get 2 confirmations (well maybe 20>2x confirmation>10)

making block size bigger really does nothing significant.  it would be fairly easy to calculate a new diff and reward for 24 sec blocks based upon one point in time.

so is 24 sec block time less secure than 600 sec block time?  (with 400 ph network?)

What is downside to shorter block times?

The idea is that it will speed up confirmation times because then the transaction won't be stuck in a backlog thus waiting even longer for two confirmations. It can pretty much guarantee that your transaction will be included in the next block, instead of say the 10th block from the time the transaction was created.

Yes shorter times would speed up confirmation times and handle more transactions, but people are generally against that idea. It would require more to change than increasing the block size does. In order to change the block time and still keep everything else the same, the block subsidy needs to be changed, the difficulty will need to be reset, and a whole bunch of other stuff as well.

As for why 10 minutes was chosen by Satoshi, it was probably arbitrary, just like basically every other number in Bitcoin.

OK, I get the idea about larger blocks = more transactions/block
more tx/second

Can someone please explain how this will help speed up a transaction requiring 2 confirmations?  I cant ever imagine anyone swiping their bitcoin card at the gas pump and waiting 20 min for "authorization"  Imagine fast food drive thru line or the checkout line in grocery store If everyone had to wait 10 min for "authorization/confirmation?

All these bigger blocks seem to be able to do is hold more dust/spam transactions.  I dont think any of this will help increase mainstream use of bitcoin.

I guess if we had a centralized online wallet that everyone used then this centralized wallet could make transactions instant.  Like I pay dish with my coinbase wallet instant confirmation.  But that goes against decentralization.

Then there is all this lightning network sidechain stuff, maybe there is a solution in the works.

So I have to ask, why do we have to have 10 min blocks?   Is there more security having 10 min vs 10 sec blocks?  Or just so the block reward is 25btc?   If mining revenue is 150 btc/hr, then why would it matter if it is 25btc/10 min block or 1 btc/24 sec block?

So, would the block size increase from 1mb block to 8mb,   be the exact same as keeping the block size the same 1 mb but reducing the time from 10 min to 75 sec?  (10*60/8=75sec)  (block reward = 25/8=3.125)

I think this would handle more transactions (or same as 8mb blocks)  AND SPEED up confirmation times. 

Why is bitcoin so married to 10 min blocks?  Is it tied the same way to 1 mb block size?

And really, will this larger block size help take bitcoin mainstream?

I know my reasoning is missing something, so I write to ask What?