Quote from: JorgeStolfi on August 24, 2015, 06:47:25 PM

Quote from: dreamspark on August 24, 2015, 06:23:08 PM

Other than the spam attack average block size at around 0.4 MB. Absolutely no where near full blocks

If you look at the last 2 years you will see that the daily average it has doubled in the last 12 months. Even if the growth is linear rather than exponential, if it continues at that pace it will be 0.60 MB/block by mid-2016.

This proves the point that there was no rush to get ...20MB or 8MB blocks.

Quote

Moreover, a "spam attack" now is very cheap because the attacker only needs to generate more than the average free capacity, that is 0.35 MB/block, to create a permanent traffic jam. By mid-2016, he will need only 0.15 MB/block.

That's more like a DOS attack that is temporary and can be circumvented by someone paying higher fees to get confirmed immediately.

Spam attack, in the sense of bloating the blockchain for the lolz, would be far cheaper with larger blocks as no competition means that a spammer could easily push through a few megabytes, with almost zero fees. And that kind of spam attack leaves a permanent imprint for the future. And then there is also the possibility that some third party actually wants to spend money in fees just to spam the blockchain for years worth of transactions, in a couple of weeks.

Quote

The blocks will not get 18 times bigger with BIP101. The will continue to be 0.40 MB now, and will continue to grow at the same pace as they would with 1 MB limit. The effect of BIP101 will start to be felt only sometime in the first half of 2016, when traffic will get to 0.55 MB/block close to the 0.75 MB capacity. With the 1 MB limit, there would start to be recurrent jams and long waits; whereas, with the 8 MB limit, the traffic will just continue growing beyond that value, at its natural rate, and maybe reach 1 MB/block in 2017.

The bolded part is an assumption. Assumption is the mother of all fuckups, and, when network abuse is concerned, it will be a guaranteed disaster.

Security assumptions like "noone is going to spam the blockchain" cannot be taken for granted. If they do, then bitcoin XT is DOA. Such a level of irresponsibility and wishful thinking is unacceptable for a multibillion dollar project that wants to rival the traditional money system.

The attack vectors surrounding a system like Bitcoin, from a game theory perspective, are wider than the Bitcoin ecosystem itself.

For example, from a game theory perspective, and if we look at this only from inside the ecosystem, a spam attack that costs a few BTC per day to make the blockchain bloated doesn't make sense, so it's illogical and would not happen.

But if we look at it from the broader financial ecosystem, a rival (a bank, a credit card, the governments etc etc) might want to do just that and be willing to pay the few BTCs in fees to bloat the hell out of the blockchain. And make sure that they do that as cheaply as possible, and with as little pruning potential as possible, by designing the proper "transactions". So they pay a few hundred BTCs per day in fees just to make the blockchain unusable.

In what financial world is it acceptable for a multibillion project to be under threat through such cheap attack vectors? Roll Eyes

The alternative attack vector of DoSing the transactions confirmed by fee competition is far more acceptable because it doesn't leave a permanent imprint and it can be bypassed by the user who wants transaction priority.

If user A wants to pay user B 10 BTCs and the confirmation can be done in one block with 0.002 instead of 0.001 btc, if he so decides to pay them, he'll pay them and get on with it.

Block size should increase in parallel with actual usage to prevent the bloat-vector conducted by a Bitcoin enemy. Scaling should be primarily improved through work in the underlying protocol or data storage.

Fee measures should already be taken to eliminate dust and spam due to this:

Quote

Quote

Most of the space is spam and dust anyway so even if we had full blocks the correct fee would get your transaction through no problem.

That is true. If a minimum fee of (say) 0.001 BTC = 0.20 USD was imposed on each transaction, the traffic would probably drop by 80% or more, and the block size limit would not be a pressing issue for a couple of years, at least. But neither camp wants to see hat, for several reasons...