Possible attack: Difficulty DoS

[Sukrim]

I'm not sure if this attack scenario already came up:

If someone posesses enough computing power to push the difficulty to a level that is far beyond the current level, this might cause the network to collapse. It might be nearly impossible after the attacker leaves the network again 2016 blocks later to discover/solve any block, so the next 2016 blocks might take ages or be never found if miners give up their business. So - is a difficulty manipulation attack possible in bitcoin (theoretically) - or are there some countermeasures that could be taken?

Just imagine for example "back in the days" when you could only do CPU hashing if somebody would suddenly drop in with a few dozen GHashes/s, solve lots of blocks in nearly no time, push the difficulty to 50 000 and disappear again. With 3 MHashes/s this would on average mean to get a hit about every 2-3 years. Even if the whole network has ~1 GH/s even a simple transaction would need 1 week to be confirmed.

In an early adopters/enthusiast community this might be ok, but later on (as it is expected somehow that difficulty might fall again when the income from BTC generation gets less and less and the few bitcents from transaction fees get more and more interesting) this could pose a real threat to the network.

[1715377547]

1715377547

[1715377547]

1715377547

[1715377547]

1715377547

[koin]

in nearly no time, push the difficulty to 50 000

the diffiulty level can only increase  4X the current level in any one period.  so instead of every 10 minutes per block, you'ld have to wait 40 minutes per block to confirm.  problematic, but not fatal.

bitcoin is between 1.2 and 2.0 thash/s currently.  good gpu hardware is sold out nearly everywhere yoo look.

what type of computing capacity do you see this attacker being able to scrounge up to be able to wage such an attack?

[theymos]

If we really got stuck that badly, a new version of Bitcoin would be released that changed the difficulty rules. I doubt it will ever be necessary, since an attacker would need a ton of computational power, and the build-up of transaction fees would incentivize increased mining.

[forgotmypassword6x]

this is a good thing; if a single person has enough power to do this, they have way more than 50% of collective hashing power in their control.  So, it would be unwise to continue to run bitcoin

[Sukrim]

bitcoin is between 1.2 and 2.0 thash/s currently.

Currently you can buy a GPU and be "quite" sure that it'll pay off in a not too distant future (at least sooner than it's expected lifetime), as there is a real run for buying BTC, which keeps the BTC<-->USD price high.
This will change over time, as the profit margin from mining will become smaller and smaller.

If you can only earn less than 1 BTC per block in the future, Bitcoin would either rather have prices comparable to gold/diamonds than to USD or hashing power/difficulty will be much lower (in proportion to future computing devices that is of course).

It's nice to hear though, that difficulty can only increase 4-fold, not skyrocket endlessly after just 2016 blocks!