nothing to see here, move along.

[RoadTrain]

   In a future bitcoin economy where most of the revenue is from tx fees this won't be an issue either, unless the blocksize is kept the same. I like this post though, becuase it creates a good case against fee contention (which is created by limited blocksize). Increasing the blocksize/removing the cap entirely solves this problem (and many others) permanently.

It also creates/worsens other problems. This has been discussed on the dev mailing list countless times.

Moreover, by dropping the cap you don't really kill the fee market, just change it. Including additional transactions in a block has its cost, e.g. increased orphaning probability.

I am personally in favor of some flexible cap that is adjusted based on actual transaction volume.

[1714854280]

1714854280

[1714854280]

1714854280

[1714854280]

1714854280

[johnyj]

Suppose that there are 10 blocks, each with a fee of 2 btc (in future). Now a large mining pool commanding 50% of hash power decided to spam the network and raise the fee for each block to 3 btc. So he spend 1 bitcoin at each block as fee for his spamming transactions, that's 10 bitcoins total, and he mined 5 of them back, his net loss is 5 bitcoins, how would he become profitable?

[johnyj]

Lets imagine your mine with half hashpower. Lets imagine that a block can contain 6000 transactions.  Attacker has 1/2 hashpower.  Offered load is 4000 tx/block.

Attacker crafts 2000/tx block at 1coin/tx fee level. Making the rest match him (plus episilon, which we'll disregard).

His average cost for spam is 1000 coin/block (2000 * 1-rate).
His average income is 2000 coin/block (4000 * rate).  (He doesn't get income from his spam, he saves its cost however; see prior line)
His net income is 1000 coins/block, on average.

Is that possible?

His average cost for spam is 1000 coin/block, but his average income is not 2000 coin/block. Because even there are 2000 transactions with a fee of 1 BTC stay at the top of the queue and get transacted immediately, the rest 4000 transactions might still pay a fee of  0.01 bitcoin, they were just pushed back in the queue, not obliged to follow the highest paying fee. The fee would only rise to 1 coin/tx if all those transactions are willing to pay for that. In a word, the spammer can only raise the fee to the amount of money he spent, the rest is not up to him

And of course he could spend huge and spam all the 6000 transactions with 1 coin/tx to occupy all the block space, then as you said that will only benefit the non-spamming miners, and those miners would still include the transactions that they want to broadcast in their blocks with a minimum fee (like exchange-mining farm cooperation)

[Taras]

Lets imagine your mine with half hashpower. Lets imagine that a block can contain 6000 transactions.  Attacker has 1/2 hashpower.  Offered load is 4000 tx/block.

Attacker crafts 2000/tx block at 1coin/tx fee level. Making the rest match him (plus episilon, which we'll disregard).

His average cost for spam is 1000 coin/block (2000 * 1-rate).
His average income is 2000 coin/block (4000 * rate).  (He doesn't get income from his spam, he saves its cost however; see prior line)
His net income is 1000 coins/block, on average.

Well, what if the attacker's block is orphaned, and other pools pick up the now-unconfirmed transactions with 1 BTC fees? Would the attacker be able to prevent that?

[DooMAD]

I'm assuming since the original post and thread title has been edited, it turns out this wasn't the case?  Even if the initial conclusions may have been a little premature in this particular instance, it's still reassuring to know that people spend time considering all the possibilities and doing their best to make sure this system continues to function as intended.  I applaud the OP's enthusiasm, but just suggest tempering it a touch and having others check the figures before announcing your findings.  Don't let this prevent you from asking bold questions in future, though, it could well pay off next time.

[turvarya]

I'm assuming since the original post and thread title has been edited, it turns out this wasn't the case?  Even if the initial conclusions may have been a little premature in this particular instance, it's still reassuring to know that people spend time considering all the possibilities and doing their best to make sure this system continues to function as intended.  I applaud the OP's enthusiasm, but just suggest tempering it a touch and having others check the figures before announcing your findings.  Don't let this prevent you from asking bold questions in future, though, it could well pay off next time.

The only premature thing, was to delete his first post.
There is nothing wrong with posting a theory. The whole point about posting a theory is so people can look into it and tell if the author has made some errors.
Nothing to be ashamed of.

[cypherdoc]

So a block size limit increase would fix the problem because for this to be profitable, the attacker needs to have at least half of the block be legitimate transactions and that there still be a backlog of transactions, yes? And since there isn't that much real transaction volume, raising the limit to the suggested compromise of 8 Mb would solve this issue for now since it suddenly becomes unprofitable to continue to attack until transaction volume increases.

Good post, totally get the math... all in on the theory...

Looking at the actuals, I'm not convinced that this is actively happening, though I see how you theorize it could.

Checking block shares, and the biggest fish only holds 20%

source: https://chain.so/btc

Though I will concead that it may be possible that the smaller pools are actually the larger holders running "shadow miners" that appear to be anon, but are really wholly owned subsidiaries of the bigger players.

Another problem I see is that blocks are not full by any stretch.  Miners (for reasons that are their own) are leaving many blocks 2/3 empty.

source: https://tradeblock.com/blockchain/

So any miner that had less than 38% the share could simply start filling blocks to the brim.  Perhaps I'm on a tangent here, but I've never understood why miners running rooms full of ASICs don't devote a few high power CPUs (useless for mining) to doing block stuffing.  It seems like getblocktemplate and hashing would be non-intersecting.  So pooring $$ into CPU could be independent to pooring money into hashing.

I still have questions tho, like how high can they push transaction fees and dose this mean the end of micro transactions in the long term?

If I followed, the condition they are driving is not the fee price, but rather the spam ratio.  As long as they can continue to mine some percentage X of their own spam then they make a profit even if fees go to 1 satoshi.  The "canaibalism" ration X would float depending on two factors... what their current share of hash rate is, and how much normal traffic there is.

but this graph from Tradeblock shows that more recent avg block sizes are just over 700kB:

[cypherdoc]

So a block size limit increase would fix the problem because for this to be profitable, the attacker needs to have at least half of the block be legitimate transactions and that there still be a backlog of transactions, yes? And since there isn't that much real transaction volume, raising the limit to the suggested compromise of 8 Mb would solve this issue for now since it suddenly becomes unprofitable to continue to attack until transaction volume increases.

Good post, totally get the math... all in on the theory...

Looking at the actuals, I'm not convinced that this is actively happening, though I see how you theorize it could.

Checking block shares, and the biggest fish only holds 20%

source: https://chain.so/btc

Though I will concead that it may be possible that the smaller pools are actually the larger holders running "shadow miners" that appear to be anon, but are really wholly owned subsidiaries of the bigger players.

Another problem I see is that blocks are not full by any stretch.  Miners (for reasons that are their own) are leaving many blocks 2/3 empty.

source: https://tradeblock.com/blockchain/

So any miner that had less than 38% the share could simply start filling blocks to the brim.  Perhaps I'm on a tangent here, but I've never understood why miners running rooms full of ASICs don't devote a few high power CPUs (useless for mining) to doing block stuffing.  It seems like getblocktemplate and hashing would be non-intersecting.  So pooring $$ into CPU could be independent to pooring money into hashing.

I still have questions tho, like how high can they push transaction fees and dose this mean the end of micro transactions in the long term?

If I followed, the condition they are driving is not the fee price, but rather the spam ratio.  As long as they can continue to mine some percentage X of their own spam then they make a profit even if fees go to 1 satoshi.  The "canaibalism" ration X would float depending on two factors... what their current share of hash rate is, and how much normal traffic there is.

but this graph from Tradeblock shows that more recent avg block sizes are just over 700kB:

yeah, the table you put up is wrong.  it shows the max block size to be 776kB which is clearly wrong.  we've had a consistent stream of 950kB+ come thru over the last few weeks.