Pages:
Author

Topic: ... (Read 1312 times)

hero member
Activity: 770
Merit: 509
...
June 24, 2015, 09:02:40 AM
#23
But we had people here yesterday complaining about some of their transactions didn't go thought for 2 hours. How's that OK?
Not to mention blockchain.info wasn't  accessible for a while.

Non were lost, if we are only at ~4TSP @ 1mb, increasing the block size is just not a solution.

We need 100 000 TSP...


IMO bitcoin will become a holding asset used to save wealth.  Something like dogecoin or bitshares will become the spending crypto people use to buy coffee / etc.



Nope, we should aim for a global Bitcoin, it's the only way to increase the price. Also if banks and NASDAQ and so on start using the blockchain to run their infrastructures we better be able to process tons of transactions per second. We need higher blocksize + LN to defeat Visa, Mastercard + Paypal in the future. Anything else but Bitcoin as the #1 electronic payment is admitting defeat.
Q7
sr. member
Activity: 448
Merit: 250
June 24, 2015, 08:55:14 AM
#22
I doubt 20 bitcoins allocated for the test can actually put out a good representation. What if the stress test prolongs and more and more backlogs start to accumulate? We are already at point where a few minutes of delay is considered slow, so how about twice of the time added to that? No matter what the outcome, I still believe increasing blocksize is one of the solutions.
legendary
Activity: 1344
Merit: 1000
June 24, 2015, 07:05:23 AM
#21
larger block sizes amoung other changes one would expect would eventually be nessesary

the only reason they arn"t already is because alts can supplement for now
hero member
Activity: 593
Merit: 500
1NoBanksLuJPXf8Sc831fPqjrRpkQPKkEA
June 24, 2015, 06:43:32 AM
#20
Thats great. I awaited to see the results.

Though one has to admit that an average natural amount of transactions exceeding 1MB would be a problem. So something has to be done with Bitcoin Core. I cant understand the way things are extorted into action.
copper member
Activity: 1498
Merit: 1528
No I dont escrow anymore.
June 23, 2015, 03:17:55 PM
#19
I would be happy to create some test transactions on the testnet.  That is what testnet is for.  Such a test should be public and it should be with the participants publishing how many transactions didn't make it and on which wallets they were using for the test transactions.

Which wallets increase transaction fees when they are not getting confirmed?  I am under the impression that a transaction gets confirmed or it sits in the network's memory.  Unfortunately I don't see how I can get electrum to use testnet, and it is probably used more than bitcoin-qt.

The Schildbach app has a testnet version and IIRC it has a fixed (or 3 step) fee setting. Thus it would be well suited to simulate a wallet that is unable to change fees depending on the network load. I can raid some testnet faucets in order to get 1-2 BTC in the apps wallet.

We know the network didn't die at 1 MB filled blocks though they filled up.  The OP didn't publish how many transactions were never confirmed or how his wallets reacted because presumably the hacker never told him.  

If all the transaction fees are 0.005 and there are 150 B per transaction, then we have 33 BTC per block from fees alone for the miners.  This seems economically healthy (for them).

sdp

My guess is that mostly the fees of the person doing the stress test did raise. Maybe a few that used bitcoin core durring that time did also pay a higher fee. The majority of users however do not use bitcoin core anymore and thus would have paid the same fee they always pay[1]. I know some noticed the hold up and complained here (or elsewhere), but I wonder whether this was an actual issue for many. That beeing said, Im in support of bigger blocks, mainly because I see them as vitally important for bitcoins future.



[1] I know some services adjust the fees as well.
sdp
sr. member
Activity: 469
Merit: 281
June 23, 2015, 11:19:57 AM
#18
I would be happy to create some test transactions on the testnet.  That is what testnet is for.  Such a test should be public and it should be with the participants publishing how many transactions didn't make it and on which wallets they were using for the test transactions.

Which wallets increase transaction fees when they are not getting confirmed?  I am under the impression that a transaction gets confirmed or it sits in the network's memory.  Unfortunately I don't see how I can get electrum to use testnet, and it is probably used more than bitcoin-qt.

We know the network didn't die at 1 MB filled blocks though they filled up.  The OP didn't publish how many transactions were never confirmed or how his wallets reacted because presumably the hacker never told him.  

If all the transaction fees are 0.005 and there are 150 B per transaction, then we have 33 BTC per block from fees alone for the miners.  This seems economically healthy (for them).

sdp
hero member
Activity: 882
Merit: 1000
June 23, 2015, 10:46:33 AM
#17
@doomad

You say that with a diminishing block reward miners wont have an incentive to secure the network.

Well, ixcoin is bitcoin, there is no more minting, and the network has continued to be secure.
hero member
Activity: 714
Merit: 500
June 23, 2015, 10:19:34 AM
#16
I'd really love to see a good analyzes of this stress test. I don't think, this is one. There are still too many assumptions in it.
It is not clear to me, if the "hacker" really ran out of money like OP claims. Is there any evidence for that? Are there good tools for analyzing the blockchain? Just looking at the graphs sites like blockchain.info provide isn't enough for a good analyzes.
legendary
Activity: 1148
Merit: 1014
In Satoshi I Trust
June 23, 2015, 10:10:14 AM
#15
here we go again...
sr. member
Activity: 274
Merit: 250
June 23, 2015, 09:44:07 AM
#14
Hi OP

If you monitored the network activity while this all happened you would have noticed at least 2 transactions where a 50BTC fee was attached.
The network did not 'fight back' and increased fees automatically...
legendary
Activity: 3430
Merit: 3080
June 23, 2015, 09:24:29 AM
#13
Block size must increase, but it's not the only change needed, because that on it's own doesn't scale up for billions of users. It works as a can-kicking, short term band-aid type fix. It is not the solution to the problem.

Completely agreed, I am quite comfortable with an increase to 8MB in early 2016. Additionally, we shouldn't automatically incorporate simple solutions to scale which involve continuously kicking the can down the road and avoiding creating novel and more efficient solutions. This hard fork "crisis" is really an opportunity to motivate us to be creative and perform more testing. This is a good thing to go through again in 2-6 years if we are forced to.

I don't enjoy dismissing those who aren't interested in technical matters, but people arguing just for a block size increase are not well informed enough to contribute a meaningful opinion to this debate (not you, BitUsher, I mean in general).

The problem is that it's really easy to paint the debate in over-simplistic terms: transactions per second has a simple linear relationship to the block size. So people can be induced to believe that biggerblocks == moretransactionspersecond+yay! The real picture has alot more moving parts than something so pedestrian as the blocksize/txps dynamic.
legendary
Activity: 1386
Merit: 1000
English <-> Portuguese translations
June 23, 2015, 09:16:22 AM
#12
Imagine fees of 0.005 are required when the bitcoin price is much higher... At 1000 usd that would be 5 usd per transaction... That means no more cheap transfers and bye bye advantage of bitcoin..

It still proves bigger blocks are required..
Compare that to western union, $50-100 fee and poor people use it all the time regardless. It just goes to show that Bitcoin will be affordable for all even if there is a huge amount of transactions.

The higher transaction fees are needed for the miners to sustain the network in the long term. The community can choose to either pay the security guard or give him a few pennies and he'll leave for another job.

And if you want to send a small 0.005(near to 1 dollar) to someone, paying 0.0005 means a fee of 10%.
This is far from cheap, since Bitcoin isn't supposed to be used only on giant transactions like bank transfers.
legendary
Activity: 994
Merit: 1035
June 23, 2015, 09:02:26 AM
#11
Block size must increase, but it's not the only change needed, because that on it's own doesn't scale up for billions of users. It works as a can-kicking, short term band-aid type fix. It is not the solution to the problem.

Completely agreed, I am quite comfortable with an increase to 8MB in early 2016. Additionally, we shouldn't automatically incorporate simple solutions to scale which involve continuously kicking the can down the road and avoiding creating novel and more efficient solutions. This hard fork "crisis" is really an opportunity to motivate us to be creative and perform more testing. This is a good thing to go through again in 2-6 years if we are forced to.
legendary
Activity: 3948
Merit: 3191
Leave no FUD unchallenged
June 23, 2015, 08:57:52 AM
#10
I see a lot of insults, but little evidence that actually tears my argument apart. Notice how I haven't insulted you once, yet you say I'm difficult to follow, laughable, a conspiracy theorist, completely backwards, cant be taken seriously, unduly panicked, unsubstantiated, fallacious.

Maybe we could focus on the actual discussion itself instead of turning this into a personal thing? I don't enjoy fighting people on the internet.

I do agree that my posts earlier that Bitcoin XT is directly related to the hacker need more proof before things like that can be said. Notice how I didn't discuss that in the article at all.

As for whether this guy is a hacker, this is the definition of hacking. Also a lot easier than saying "the person running the stress test" every time I reference him.

The only thing I don't like about this hacker is that they are hiding the results and trying to tell the community there will be a hack 10X bigger in a week. Fairly immature tactic, the chart above is completely inconsistent with their story that bitcoind crashed early in the test. 18 hours is a decently long time and there were multiple spikes of high transaction fees that evaporated their Bitcoin supply.

Let's run through some of those points, then.

If you'd like to explain how your logic isn't contradictory, I'd love to hear it.  Again, if this short term period of disruption is "inconsiderate", "malicious" and "selfish", how is risking the network frequently running in such a fashion by imposing an artificial bottleneck somehow a reasonable option in your opinion?  You can't have it both ways.  If you think you can, then your logic is difficult to follow.

Jumping to the conclusion in the other thread that this test is tied to the XT developers was unsubstantiated and does sound like the kind of conclusion a conspiracy theorist would jump to, but if you're now withdrawing that comment, I'll happily rescind mine.

There's no other way to describe your argument about fees, it is completely backwards and I feel I've substantiated that claim in the quoted segments in my prior post.  If the network never supports more than a set, finite number of transactions per block, even if those users paid a higher fee, there will be insufficient incentive for miners to secure the network once the block reward starts to diminish.  You could possibly gamble on the price of a single BTC rising, but that's a major gamble if we start to shed users as a result of poor service because their transactions get stuck in a queue.  Bitcoin currently maintains a reputation for being fast, reliable and cheap, but if we ever lose that, I don't see how it could possibly bode well for the price.

Your statement that "Limited block size would absolutely not limit the amount of users, everyone can send as much transactions as they want at any time as long as they pay a fee" is completely, unequivocally and undeniably fallacious.  It's just plain wrong to state that as a fact when it's untrue.  Once a block is full, by definition, it is full.  At present, you can't fit more than 1MB worth of transactions into one block.  Ergo, if there are more than 1MB worth of transactions waiting, some of those transactions will be delayed, even if a fee is included with every single one.  There is literally no denying this.  That's how Bitcoin works at present.  If nothing else, you need to concede that point.
legendary
Activity: 3430
Merit: 3080
June 23, 2015, 08:53:42 AM
#9
Imagine fees of 0.005 are required when the bitcoin price is much higher... At 1000 usd that would be 5 usd per transaction... That means no more cheap transfers and bye bye advantage of bitcoin..

It still proves bigger blocks are required..

Block size must increase, but it's not the only change needed, because that on it's own doesn't scale up for billions of users. It works as a can-kicking, short term band-aid type fix. It is not the solution to the problem.
legendary
Activity: 3430
Merit: 3080
June 23, 2015, 08:49:31 AM
#8
As for whether this guy is a hacker, this is the definition of hacking.

No it isn't. Hacking is the computer networking equivalent of breaking and entering. I though you were arguing against the use of ad hominem attacks misrepresented as argument?


legendary
Activity: 1946
Merit: 1007
June 23, 2015, 08:45:50 AM
#7
Imagine fees of 0.005 are required when the bitcoin price is much higher... At 1000 usd that would be 5 usd per transaction... That means no more cheap transfers and bye bye advantage of bitcoin..

It still proves bigger blocks are required..
legendary
Activity: 3948
Merit: 3191
Leave no FUD unchallenged
June 23, 2015, 08:05:18 AM
#6
*sigh*

You posted these same conclusions in that thread, so unless you were trying to conveniently gloss over the part where several users tore your argument to pieces, what purpose was there in creating another thread to say exactly the same thing?  I suspect that yours will be another one of those accounts that gets added to my ignore list after the fork.  I'd block you now, but if left unchallenged, other users might mistakenly believe your "findings".

Here's some of the bits you left out:

but doing this sort of thing is extremely inconsiderate to Bitcoin users. Even if the intentions are to collect data it's still a malicious attack and hurt the Bitcoin economy today, very selfish.

Once again, your logic is proving difficult to follow.  If this short term period of disruption is "inconsiderate", "malicious" and "selfish", how is risking the network frequently running in such a fashion by imposing an artificial bottleneck somehow a reasonable option in your opinion?  Obviously these weren't important transactions because it was a test, so it's inconsequential that they weren't able to continue transacting.  But there will likely come a time when more people start to use Bitcoin for normal transactions.  If the userbase is sufficiently large enough, what we've seen in this test would be the tip of the iceberg.  It will be regular users that are forced to stop sending transactions because of delays or unexpected rises in cost and this will be far more damaging to Bitcoin's longevity than any supposedly malicious test.  If Bitcoin doesn't scale, the users that can't be accommodated comfortably will simply switch to a more convenient payment method.  One that doesn't have varying fees and unexpected delays.


The hacker had the same belief as the XT team that Bitcoin is in imminent danger from the blocksize limit, and spent a large sum of money to try and force the situation. There is no direct proof yet though.

Firstly, "hacker"?  Seriously?  Your attempts to discredit anyone who disagrees with you are laughable.  Secondly, anyone with eyes can see that the blocksize limit is an issue, otherwise there wouldn't be so many threads discussing it.  Not everything is some bizarre conspiracy. 


In fact, I think this proved that Bitcoin naturally responds to such an attack/transaction volume increase naturally and eliminates the problem. Transaction fees simply go up so people stop sending spam/dust transactions and only important transactions. This suggests a blocksize increase isn't necessary.

Just understand that once there are more users, it won't just be spam or dust transactions, it was be all transactions that are subject to these conditions.  You will be paying the extra fees when the network is busy and if people pay a higher fee than you, it's your transactions that might be delayed.  We've never run the network in that fashion before, but if you want to attempt that, then it's your decision to vote that way if you mine or run a node.

So you completely fail to see my point that transaction fees will be the only thing keeping mining alive in the future? If we went along with XT Bitcoin would die as block rewards become tiny.

Nope, got it completely backwards again.  If there were no block reward right now, miners would have little incentive to secure the network with the amount of fees currently collected. 

If, hypothetically, there was no block reward right now, based on the transaction volume we are getting now, how much fee would need to be paid on each transaction to give at least the current average level of fees plus the 25 BTC reward?  Last time I bothered to do the math, we were averaging about 750 transactions per block.  By my figures, each transaction would have to have a fee of about .034 BTC just to cover the 25 BTC reward, plus a bit extra for the current fees that are already being paid (so there's no loss of mining revenue).  How many people are going to pay about $8 or $9 USD in fees for every single transaction?  Most would switch to an altcoin in a heartbeat if Bitcoin became that expensive.  If people aren't willing to pay that much, there will be even fewer transactions and that cost will then rise even further.  But increase the number of transactions in the block, rather than limiting it, then you can spread that cost over a greater number of users, making it more affordable.  Obviously we can squeeze in a few more than the 750 we're averaging at the moment, but we genuinely do need more users than the system will currently support to make this thing sustainable in future.  Unless you foresee Bitcoin becoming a niche, elitist, isolated haven for wealthy one-percenters, there is simply no other option than to accommodate more users.  More users equals more fees for miners, not less.

Limited block size would absolutely not limit the amount of users, everyone can send as much transactions as they want at any time as long as they pay a fee. Even if it's a few dollars that's still pretty good, especially if bitcoin is well over $1000 by then, and it damn well should be.

Categorically 100% false.  The 1MB limit is currently a hard limit, so by it's very definition, it limits that number of transactions the network can process.  If, for example, there were 20MB worth of transactions waiting, even if everyone paid over $10 dollars in fees, the network is unable to process them all in one go.  It would take 20 blocks to clear them with a 1MB blocksize.  That's likely to be over 3 hours for some of those transactions.  Would you not be annoyed if you paid a fee and still found you had to wait 3 hours because other people had paid a higher fee?  However, with an 8MB blocksize, you will still pay a fee to prioritise your transaction, but we could clear that queue in three blocks.  On average you'd be waiting 30 minutes at most.  This is the part most people don't seem to comprehend.  As the userbase grows, the blocksize must grow to accommodate it.
This is the same flawed thinking that led to Bitcoin XT. Confirmation time will always be on average 10 minutes for transactions with the proper fees.

I can assure you there is no flaw in what I've said, I've simply described how the protocol works.  You're the one who fails to understand it correctly.  The confirmation time for some of the transactions will be on average 10 minutes.  But if there are more than 1MB worth of transactions waiting, even if every single one of those transaction has a fee included, some of those transactions will have to wait for another block with space available if we stick with a 1MB limit.  If that still doesn't make sense to you, then I'm afraid I can't take your arguments seriously.  It's no wonder you're unduly panicked by the whole situation if you haven't taken the time and effort to understand why the fork is being proposed.  If you continue to post unsubstantiated accusations, wild conspiracy theories and outright fallacious statements, it's unlikely anyone else will take you seriously either.
legendary
Activity: 994
Merit: 1035
June 23, 2015, 07:49:01 AM
#5
But we had people here yesterday complaining about some of their transactions didn't go thought for 2 hours. How's that OK?
Not to mention blockchain.info wasn't  accessible for a while.

It isn't ok, Peter Todd has some potential solutions -

https://gist.github.com/petertodd/8e87c782bdf342ef18fb

Quote from: Peter Todd
What we need to do:

    Add fee/KB displays to block explorers.

    Change wallets to calculate and set fees in fee/KB rather than fixed fees regardless of tx size.

    Make websites with easy to understand displays of what the current mempool backlog is, and what fee/KB is needed to get to the front of the queue. We've done a great job for Bitcoin price charts, let's extend that to transaction fees.

    Add the ability to set any fee/KB to wallets, rather than be stuck with predefined options that may not be high enough.

    Add support for fee-bumping via (FSS)-RBF to wallets and Bitcoin Core.

I still think that we need to raise the blocklimit to 8MB to give other solutions a bit of breathing room , but both Garziks and Gavins proposals still need more adjustments and testing.
legendary
Activity: 3430
Merit: 3080
June 23, 2015, 07:46:01 AM
#4
But we had people here yesterday complaining about some of their transactions didn't go thought for 2 hours. How's that OK?
Not to mention blockchain.info wasn't  accessible for a while.

I sent transactions on Monday, no such problems.
Pages:
Jump to: