network security and the size of mining economy

[niko]

This has been touched upon in various threads, I'd like to discuss it in more detail.

Transaction fees are currently a negligible fraction of the miner's income, it's all about the block reward. There are presently 7200 coins scooped up every day, soon to be halved to 3600.  Assuming most miners will not want to mine at a loss in the long term, and assuming they expect a reasonable period of return on investment of say one year, they will only purchase up to 3600*365 = 1.3 million coins worth of equipment and electricity. This is the maximum value of mining equipment in the world, assuming mostly rational decisions are made. Anyone wanting to stage a disruption of the network (aka >50% "attack") will need to spend roughly that much on their mining equipment and electricity. This is the best-case scenario, where the attacker relies on best publicly available mining technology, instead of investing into development of their own technology.

All the talk about ASICs making the network more secure is irrelevant: whatever currently available best technology is (GPUs, FPGA, ASICs, ) - the attacker needs to spend an equivalent of roughly one or two million coins to stage an attack (or more precisely, a disruption).

One caveat is that some miners might be investing gambling based on expectations of future value of coins, so their equipment and electricity would cost somewhat more in fiat, but I'd be surprised if this plays a big role.

What do you think?

[Korbman]

Transaction fees are currently a negligible fraction of the miner's income, it's all about the block reward. There are presently 7200 coins scooped up every day, soon to be halved to 3600.  Assuming most miners will not want to mine at a loss in the long term, and assuming they expect a reasonable period of return on investment of say one year, they will only purchase up to 3600*365 = 1.3 million coins worth of equipment and electricity. This is the maximum value of mining equipment in the world, assuming mostly rational decisions are made. Anyone wanting to stage a disruption of the network (aka >50% "attack") will need to spend roughly that much on their mining equipment and electricity. This is the best-case scenario, where the attacker relies on best publicly available mining technology, instead of investing into development of their own technology.

All the talk about ASICs making the network more secure is irrelevant: whatever currently available best technology is (GPUs, FPGA, ASICs, ) - the attacker needs to spend an equivalent of roughly one or two million coins to stage an attack (or more precisely, a disruption).

The only way this would be possible is if someone was actively attempting to destroy Bitcoin. By creating a "51% attack", pretty much every user would dump their coins for normal currency, crashing the price, and essentially making Bitcoin worthless. There's no point in mining if someone can control the network.

So it's great if some company wanted to spend a few $million on equipment to destroy the system, but once people stop using Bitcoin all that equipment would be useless as well.

[niko]

Good, we agree there. Is it fair to put it this way then: right now Bitcoin is relatively obscure, and no entity is likely to want to spend BTC1.3M * 10 $/BTC to disrupt the network. If and when Bitcoin grows big enough for someone to care to disrupt it, it will be much harder to do it,  because value of BTC1.3M will be much higher.