Next level Bitcoin stress test -- June 29-30 13:00 GMT 2015

[RustyNomad]

Just read through must of this thread and I think we have it the wrong way round....

Seems like OP is not doing a stress test on the blockchain but rather on their 32 servers and clearly they cannot handle the stress...

On a serious note though. I think this test was a good thing overall, even if it did not go according to plan. For me personally it helped clarify many things in regards to the 'increase/don't increase the block size' debate.

[1713301930]

1713301930

[1713301930]

1713301930

[AgentofCoin]

No that's still the backlog from the short 15 minute long spike of the actual test.
Transaction number is still slightly elevated but FAR from what KingAfurah proposed for their test.
Check: http://statoshi.info/dashboard/db/transactions
You can see the spike of the actual test that occurred hours ago, but did not last very long.
KingAfurah care to explain what's going on?

I'm not sure that website is correct. Either it is buggy or something is going on.

2015-06-30 6:54:30 = 2153 tx in mempool
Then 10 minutes later
2015-06-30 7:04:30 = 13,760 tx in mempool

That is a 11,500 tx increase in 10 minutes. That isn't a slow build, that is some error, IMO.

[cryptrol]

Maybe the people behind this "tests" (obviously lobbying for a block size increase) have decided that it is better to do a lower intensity transaction flood during a larger period of time than to do a massive flood in a very short period of time. This is sad because the latter would be much more interesting IMO, I bet a mempool of 150Mb or more is not easy to handle.

[ChetnotAtkins]

Mempool at 12Mb

Maybe the test works after all  Still far below their goals though

[cryptrol]

Mempool at 12Mb

Maybe the test works after all  Still far below their goals though

There is no technical reason for the test NOT to work, the non obvious part would be the OPs agenda and motivations to do it.

[ChetnotAtkins]

Transactions per second are not unusual though. I wonder what's going on on their end that prevents them from sending the 32 tx/s as planned.

[sturle]

And what happens after all the "spam" transactions are delayed and adoption has grown to a point where there still isn't enough room in a 1MB block for more "normal priority" fee paying transactions?  That's right, the normal priority transactions will start to be delayed as well.  What do you propose then?  Or do you honestly believe roughly 10 minutes worth of transactions from all over the globe will continue to fit neatly into something smaller than a floppy disk for the rest of forever?  

People will probably send smarter transactions if the fee gets too high.  E.g. joining more transactions in sendmany, using internal transactions at exchanges where both have a wallet, etc.  Bitcoin will not go under, and Bitcoin transaction fees have lots of room for expansion before we reach credit card or bank transfer level.  Eventually we will have to use sidechains, Lightning or similar for smaller transactions where the fee will be too high for economic use of an in-chain transaction.  And we will get faster transactions as well.  But there is lots of room for growth before we need that, as this malicious spam campaign has proved.

[ChetnotAtkins]

Impressive a 15-30 min burst of spam transactions has backlogged for so long, every block has been filled for the past 5 hours.

I agree. Imagine what an extended period of increased transaction count would have caused!

I am still observing, although I have to admit that I still believe this test, for whatever reason, is still a failure in regards to their actual plan. We won't know anything until they comunicate again. Maybe they'll manage to increase tx/s to 32 again, not sure what's stopping them.

[DooMAD]

And what happens after all the "spam" transactions are delayed and adoption has grown to a point where there still isn't enough room in a 1MB block for more "normal priority" fee paying transactions?  That's right, the normal priority transactions will start to be delayed as well.  What do you propose then?  Or do you honestly believe roughly 10 minutes worth of transactions from all over the globe will continue to fit neatly into something smaller than a floppy disk for the rest of forever?  

People will probably send smarter transactions if the fee gets too high.  E.g. joining more transactions in sendmany, using internal transactions at exchanges where both have a wallet, etc.  Bitcoin will not go under, and Bitcoin transaction fees have lots of room for expansion before we reach credit card or bank transfer level.  Eventually we will have to use sidechains, Lightning or similar for smaller transactions where the fee will be too high for economic use of an in-chain transaction.  And we will get faster transactions as well.  But there is lots of room for growth before we need that, as this malicious spam campaign has proved.

You think the average person will put up with all that needless extra crap?  I highly doubt it.  People like simple and they like fair.  The only reason to limit the network to 1MB is to pander to elitist ideals.  People sending large amounts get security, everyone else gets risk and is forced off-chain to rely on a third party.  The wealthy minority get all the benefit and everyone else gets a second tier service.  People getting involved in crypto aren't doing so just to rely on a different third party to handle their transactions for them.  The whole idea is that the transaction is recorded on a blockchain without placing trust in someone else.  If Bitcoin can't accommodate those transactions, the majority will simply use a different cryptocurrency.  Users won't support a network that doesn't support them.

If Bitcoin becomes nothing more than a tool for the wealthy, I don't see how they're going to keep miners on board.  Billions of people sending several transactions a day on a network that does scale will easily generate more in fees than a few one-percenter-wanabees sending the occasional large payment.  As Bitcoin's block reward diminishes over time, those fees will need to increase if you don't want to start haemorrhaging hash rate.  The question is, do you make individual users pay a higher amount?  Or do you allow the number of transactions to increase and collect a greater quantity of fees.  

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.  

[Alley]

So servers crashed again?  They might want to do a stress test on there own systems first.

[ChetnotAtkins]

Ok it's getting interesting again. According to Statoshi we are currently at 17 tx/s !

[Quickseller]

Ok it's getting interesting again. According to Statoshi we are currently at 17 tx/s !

Yea, if you look at blockchain.info's main page, you will see new transactions scrolling by so quickly that you can't even see any of them.

Total fees of unconfirmed transactions is currently ~2.04 BTC, the mem pool is ~12.25 MB, the last block was ~4 minutes ago and the 6th last block was found ~51 minutes ago

[acquafredda]

stress test + greece fear + something else = BTC UP

[sturle]

And what happens after all the "spam" transactions are delayed and adoption has grown to a point where there still isn't enough room in a 1MB block for more "normal priority" fee paying transactions?  That's right, the normal priority transactions will start to be delayed as well.  What do you propose then?  Or do you honestly believe roughly 10 minutes worth of transactions from all over the globe will continue to fit neatly into something smaller than a floppy disk for the rest of forever?  

People will probably send smarter transactions if the fee gets too high.  E.g. joining more transactions in sendmany, using internal transactions at exchanges where both have a wallet, etc.  Bitcoin will not go under, and Bitcoin transaction fees have lots of room for expansion before we reach credit card or bank transfer level.  Eventually we will have to use sidechains, Lightning or similar for smaller transactions where the fee will be too high for economic use of an in-chain transaction.  And we will get faster transactions as well.  But there is lots of room for growth before we need that, as this malicious spam campaign has proved.

You think the average person will put up with all that needless extra crap?

What needless extra crap?  I didn't mention any needless extra crap.

I highly doubt it.  People like simple and they like fair.  The only reason to limit the network to 1MB is to pander to elitist ideals.  People sending large amounts get security, everyone else gets risk and is forced off-chain to rely on a third party.

You obviously have no idea what you are talking about.  People who don't mind weaker security, already use SPV wallets instead of running a full node.  Lightning will provide at least as good security as an SPV wallet, with faster confirmation. 

The wealthy minority get all the benefit and everyone else gets a second tier service.  People getting involved in crypto aren't doing so just to rely on a different third party to handle their transactions for them.  The whole idea is that the transaction is recorded on a blockchain without placing trust in someone else.  If Bitcoin can't accommodate those transactions, the majority will simply use a different cryptocurrency.  Users won't support a network that doesn't support them.

Yes, using some scamcoin could be a solution for those with very special needs.  I expect most people to use sidechains instead, where you don't place trust in anyone else.  Don't forget the fact that over the past year or so 94% of the full nodes have left the network.  The quickly increasing resource requirements is an important reason.  I have switched off two of three of my own wery well connected full nodes for that reason myself.  The security and privacy of a full node is already a privilege of the comparatively wealthy.  You want to make it even more so.  I disagree with that approach for the many of the same reasons that you want to increase the resource requirements for privacy and security even more.

If Bitcoin becomes nothing more than a tool for the wealthy, I don't see how they're going to keep miners on board.  Billions of people sending several transactions a day on a network that does scale will easily generate more in fees than a few one-percenter-wanabees sending the occasional large payment.  As Bitcoin's block reward diminishes over time, those fees will need to increase if you don't want to start haemorrhaging hash rate.  The question is, do you make individual users pay a higher amount?  Or do you allow the number of transactions to increase and collect a greater quantity of fees.

You don't get it, do you?  Increasing fees means increasing miner rewards at no cost to the miners.  Larger blocks incurs a cost, which is why many miners don't fill the blocks even today.

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?

What an incredibly useless exercise.  When I started the block reward was 50 BTC.  There was a neglible drop in hashrate when the reward halved.  There is no reason to keep the reward at 25 BTC.  The cost of that to everyone who want the security and privacy of running a full node is too high.

[sturle]

Ah, the drama.  Looks like this malicious spam attack has no effect on other transactions than spam as well.  From #bitcoin on Freenode:

19:30 <@wizkid057> xx1d: I've been filtering the attack on Eligius since the first one (as you'll notice Eligius does not mine the attacker's spam).  I've also been monitoring Eligius's memory pool size.  As other pools mine spam filled blocks they also are mining the majority of the legit txns that Eligius is trying to mine, dropping Eligius's memory pool down to just tens of KB... meaning the attack is
19:30 <@wizkid057> effectively ineffective even without Eligius explicitly filtering (the filtering just gives me a good metric to monitor)

Just as the previous malicious spam attack.  An effective attack is going to cost a lot more, and require a large amount of capital in form of older larger inputs and larger outputs which are alowed to age before they are spent again.  The spammer is wasting his coins again, proving that bitcoin works well.  Smaller blocks only means the spammer wastes his coins more slowly.

[weex]

[DooMAD]

What needless extra crap?  I didn't mention any needless extra crap.

You said

using internal transactions at exchanges where both have a wallet

but if I know 10 different people who use 10 different exchanges, I'm clearly not going to sign up to all of them because that's somehow "smarter".

Lightning will provide at least as good security as an SPV wallet, with faster confirmation.

And the official release date for Lightning is when?  

I expect most people to use sidechains instead, where you don't place trust in anyone else.

And the official release date for sidechains is when?

(and the anti-fork crowd accuse the pro-fork crowd of kicking the can down the road - at least we have some deadlines in mind    )

You don't get it, do you?  Increasing fees means increasing miner rewards at no cost to the miners.  Larger blocks incurs a cost, which is why many miners don't fill the blocks even today.

I get it just fine, thanks.  But I'm pretty sure there will be some miners who would like to mine some larger blocks and collect the extra fees from all the additional transactions they'll be processing.  Hence several of the large pools expressing an interest in 8MB blocks.  I suppose you think they don't get it either?

There is no reason to keep the reward at 25 BTC.  The cost of that to everyone who want the security and privacy of running a full node is too high.

So you honestly expect miners to take a hit every single time the reward halves?  I thought I was the one who didn't understand their costs?  One way or another, fees need to go up.

[Quickseller]

Ah, the drama.  Looks like this malicious spam attack has no effect on other transactions than spam as well.  From #bitcoin on Freenode:

19:30 <@wizkid057> xx1d: I've been filtering the attack on Eligius since the first one (as you'll notice Eligius does not mine the attacker's spam).  I've also been monitoring Eligius's memory pool size.  As other pools mine spam filled blocks they also are mining the majority of the legit txns that Eligius is trying to mine, dropping Eligius's memory pool down to just tens of KB... meaning the attack is
19:30 <@wizkid057> effectively ineffective even without Eligius explicitly filtering (the filtering just gives me a good metric to monitor)

Just as the previous malicious spam attack.  An effective attack is going to cost a lot more, and require a large amount of capital in form of older larger inputs and larger outputs which are alowed to age before they are spent again.  The spammer is wasting his coins again, proving that bitcoin works well.  Smaller blocks only means the spammer wastes his coins more slowly.

The attack is not meant to harm the network. This test is designed to measure the effect of what will happen when transaction volume starts to pick up naturally.

Eligius filtering out the test's transactions is doing nothing but harm to Bitcoin and is a good reason why no one should mine on eligius. Filtering out transactions that belong to a certain person harms bitcoin's fungibility and it is things like their actions that will lead to Bitcoin's failure.

[bitcoinfees]

Nice animated graphs, weex.

Here's a plot of the mempool size for the past day, for comparison to normal levels.  

https://i.imgur.com/tEPlmCr.png

Also note the spike in required fees, brought about by the higher-than-average fees being paid by the spammed transactions.

Live graphs are available here.

[AgentofCoin]

Ah, the drama.  Looks like this malicious spam attack has no effect on other transactions than spam as well.  From #bitcoin on Freenode:

19:30 <@wizkid057> xx1d: I've been filtering the attack on Eligius since the first one (as you'll notice Eligius does not mine the attacker's spam).  I've also been monitoring Eligius's memory pool size.  As other pools mine spam filled blocks they also are mining the majority of the legit txns that Eligius is trying to mine, dropping Eligius's memory pool down to just tens of KB... meaning the attack is
19:30 <@wizkid057> effectively ineffective even without Eligius explicitly filtering (the filtering just gives me a good metric to monitor)

Just as the previous malicious spam attack.  An effective attack is going to cost a lot more, and require a large amount of capital in form of older larger inputs and larger outputs which are alowed to age before they are spent again.  The spammer is wasting his coins again, proving that bitcoin works well.  Smaller blocks only means the spammer wastes his coins more slowly.

The attack is not meant to harm the network. This test is designed to measure the effect of what will happen when transaction volume starts to pick up naturally.

Eligius filtering out the test's transactions is doing nothing but harm to Bitcoin and is a good reason why no one should mine on eligius. Filtering out transactions that belong to a certain person harms bitcoin's fungibility and it is things like their actions that will lead to Bitcoin's failure.

Agreed.
This also creates the precedence where miners are nonchalantly and flippantly willing to "blacklist" addresses.
A dangerous road to dance down. Eligius is complacent in the erosion of Bitcoin/bitcoin.