And if half of the miners turn off their hardwares ?

[algorista]

Inverted 50% attack. And if half of the miners turn off their hardwares ?

Anyone could please explain the consequences of an attack like this ? (Intentionally made, of course).

I can imagine an never finalized next block.

I understand the inner workings of the Bitcoin network, and I think that an attack like this could be feasible at this time, with the increasing development of ASICs.

*** Sorry if the english is not perfect, i use the google translator.

Thanks.

[mb300sd]

Avage block time would be 20 minutes for a few weeks, then back to normal.

[evansearle42]

Avage block time would be 20 minutes for a few weeks, then back to normal.

This, then our difficulty would be half and everyone would celebrate.

[gmaxwell]

No doubt the perpetrators of this "attack" can expect much riches. I recommend they commence immediately.

[hayek]

It's hard to imagine this scenario.

Displacing this many miners on the network/getting that many to corroborate is unlikely.

The difficulty would still remain at X until readjustment so it's not like the remaining miners would profit that quickly and this would just increase the amount of time the other miners would have to be down.

[balanghai]

That'd be great why not turn off 90% of the miners?

[RodeoX]

What if 90% of us deleted our wallets!!!11!! A 90% attack!     

[Buffer Overflow]

Avage block time would be 20 minutes for a few weeks, then back to normal.

This, then our difficulty would be half and everyone would celebrate.

Why? Higher the better.

[alp]

Avage block time would be 20 minutes for a few weeks, then back to normal.

This, then our difficulty would be half and everyone would celebrate.

Correction - miners would celebrate.  Those who prefer a secure network would be sad.

[Carlton Banks]

Also, previously uneconomic equipment (at the margins thereof) suddenly becomes worth mining with for the duration of the attack... the more of the network that goes down, the wider range of older equipment that becomes viable again. The peak hashrate will not be achievable (probably), but it could turn a 5 week difficulty adjust into, say, four and a half. (bitcoin) World doesn't end.

[algorista]

When I posted this question I was thinking more about the possibility of If at this moment there is already 50% control over the network, with some obscure government agency owning the majority of ASICs. I'm just imagining a scary scenario, not being a real thing.

I also still do not understood the real effects on the calculation of the difficulty when occurs an increase or decrease in hashpower on the whole network.

Assuming that 50% of hashpower is exactly 50% of the difficulty or have any variation  For example, with 50% more hashpower we have only 10% more on difficult or have an equivalence of 50% == 50% ?

Assuming that an increase in hashpower might (this is a question) rise exponentially the difficulty then when removing this extra hashpower no others will be able to meet the demand to complete a block, even in years ?

Thank you for the patience.

[riplin]

You can see the hash power distribution here: http://blockchain.info/pools

There is currently no single miner controlling 50% of the network or even close to it.

[biggie]

That is the beauty about bitcoin, decentralized ! No one must control 50 %+

[Kluge]

Avage block time would be 20 minutes for a few weeks, then back to normal.

This, then our difficulty would be half and everyone would celebrate.

50% then goes back online, making actual attacks a bit easier to pull off. Hyper50 Attack. We need more unique names for different theoretical attacks. I was thinking earlier, about the one-Satoshi dust transactions I'd guess are aimed at connecting addresses to an identity. I decided it should be called Dusty Taint Sniffing. ... Uh - what were we talking about?

[DeathAndTaxes]

When I posted this question I was thinking more about the possibility of If at this moment there is already 50% control over the network, with some obscure government agency owning the majority of ASICs. I'm just imagining a scary scenario, not being a real thing.

I also still do not understood the real effects on the calculation of the difficulty when occurs an increase or decrease in hashpower on the whole network.

Assuming that 50% of hashpower is exactly 50% of the difficulty or have any variation  For example, with 50% more hashpower we have only 10% more on difficult or have an equivalence of 50% == 50% ?

Assuming that an increase in hashpower might (this is a question) rise exponentially the difficulty then when removing this extra hashpower no others will be able to meet the demand to complete a block, even in years ?

Thank you for the patience.

No.  There is a perfect linear relationship between difficulty and hashrate.

If 50% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~20 minutes and 0-4 weeks later the network would adjust and the time between blocks would fall back to 10 minutes.

If 75% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~40 minutes and 0-16 weeks later the network would adjust and the time between blocks would fall back to 10 minutes.

If 90% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~ 100 minutes and 0-20 weeks later the network would adjust and the time between blocks would fall to 25 minutes* and then 16 weeks later it would adjust again back to 10 minutes.

If 99% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~1000 minutes and 0-200 weeks later the network would adjust and the time between blocks would fall to 250 minutes* and then 50 weeks later it would adjust again to ~60 minutes and then 12 weeks later it would adjust again to ~15 minutes ....

On edit: * network difficulty adjustment capped at 1/4 and 4x.  Of course rules can be changed and is Bitcoin is "stuck" there will be significant consenus to adjust the difficulty either as a one time change or a more adaptive difficulty algorithm.


There is no probable scenario where even a malicious removal of hashpower would result in block time taking a year or more.   1 year = 87,600 minutes.  10 minute target = 8,760x larger.  1/8,760 = 0.014%.  It would require 99.986% of the network to shutdown overnight.   Since no single entity owns 99.986% of the network, hell no contininent has that much we likely are talking an extinction level event (i.e. 99.986% of nodes go offline because they no longer exist due to an asteroid impact).

[algorista]

DeathAndTaxes, thats a perfect explanation.

Thank you.

[riplin]

If 90% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~ 100 minutes and 0-20 weeks later the network would adjust and the time between blocks would fall back to 10 minutes.

If 99% of hashpower shut off (for whatever reason) then the time between blocks would rise to ~1000 minutes and 0-200 weeks later the network would adjust and the time between blocks would fall back to 10 minutes.

Sorry, but no. The difficulty changes are capped at max 4x current difficulty and 1/4th current difficulty.

Edit:

https://github.com/bitcoin/bitcoin/blob/master/src/main.cpp

main.cpp, line 1299, GetNextWorkRequired(...)

specifically, lines 1338 ... 1341:

Code:

    if (nActualTimespan < nTargetTimespan/4)
        nActualTimespan = nTargetTimespan/4;
    if (nActualTimespan > nTargetTimespan*4)
        nActualTimespan = nTargetTimespan*4;

[algorista]

Sorry, but no. The difficulty changes are capped at max 4x current difficulty and 1/4th current difficulty.

Very interesting.

Then someone with sufficient hashpower (lets say someone with an ASIC production line, and infinite paper money) could cause some problems to the network if suddenly remove their hardwares.

Obviously such an attack would not come from someone who seeks profit in mining, but would come from someone with an interest in weakening the bitcoin markets.

[algorista]

And, of course, a Extinction Level Event could cause some problems to bitcoin too

[DeathAndTaxes]

Sorry, but no. The difficulty changes are capped at max 4x current difficulty and 1/4th current difficulty.

Good point forgot about that.  Fixed.