Quote from: johnyj on September 11, 2015, 02:54:03 PM

It just proved that bigger blocks are even more problematic during such kind of stress: The attacker can still spend 200 BTC and create 10x more traffic with 10x lower fee in each transaction, but the 10x larger mempool and 10x larger block will definitely crash most of the nodes

That's not how it works. The block size limit doesn't allow for lower fees. It only allows miners to process larger blocks if and when desired to clear out a larger number of transactions. Think of it this way: most of the time, not during stress tests, blocks are not full; the 1MB limit exists but there is still room to go as miners create smaller blocks. This doesn't change the fee people pay to have transactions relayed and included. The only difference large blocks have on a 'stress test' is they allow a miner to clear out a backlog of transactions, meaning users are less likely to experience confirmation delay. A limit ten times higher means it costs at least ten times more to create a backlog.

Quote from: adamstgBit on September 11, 2015, 03:24:04 PM

bitcoin is passing the stress test, that's what i see. if anything this is a point for 1MBers.

What would happen if the load was 10 times more intense? Right now Bitcoin doesn't have tremendous adoption, only a few million users or so, but everyone expects that to change. What happens if blocks are normally full and a test occurs, say 10 times larger? Coinwallet's budget is $50K, but it's conceivable to imagine enemies with $500K resources. Western Union, for example, profits 40 million dollars per day.