Author

Topic: If there was a 51% attack... (Read 1074 times)

legendary
Activity: 3388
Merit: 4615
June 15, 2014, 03:07:50 PM
#16
BITCOIN UNDER ATTACK!

{A whole lot of rediculout FUD}

Fucking GHASH!

What a bunch of FUD.

It is not possible in bitcoin for the blockchain to have duplicate transactions.  What the heck are you going on about?

I haven't looked at it very closely yet, but my guess is you are looking at the result of an orphan block.
copper member
Activity: 1498
Merit: 1499
No I dont escrow anymore.
June 15, 2014, 06:12:01 AM
#14
Double spends could hurt casinos that give 0 confirmation credits for deposits. It would make every casino switch to 1 confirmation, which I believe PrimeDice has done recently. Kiss

Thats BS. If I can redo 1 block I can redo 2 blocks as well. As was pointed out in on of the other threads about this topic you might only realize after 100 or more confirmations that its a doublespend.
full member
Activity: 169
Merit: 100
June 15, 2014, 01:35:55 AM
#13
Double spends could hurt casinos that give 0 confirmation credits for deposits. It would make every casino switch to 1 confirmation, which I believe PrimeDice has done recently. Kiss
sr. member
Activity: 252
Merit: 250
June 14, 2014, 12:08:36 PM
#12
What I want to contribute here, that double spending won't be a risk for us, as among the other threads here, double spending is not such a big deal.
legendary
Activity: 826
Merit: 1000
amarha
June 14, 2014, 12:00:42 PM
#11
51% of the accidents will not happen, I believe bitcoin team.

Too late Tongue

Explain?  I've not heard any news about an attack yet?

Not an attack, an "accidental" crossing of 50% is what I thought he meant.
legendary
Activity: 3388
Merit: 4615
June 14, 2014, 11:32:52 AM
#10
51% of the accidents will not happen, I believe bitcoin team.

Too late Tongue

Explain?  I've not heard any news about an attack yet?
legendary
Activity: 826
Merit: 1000
amarha
June 14, 2014, 03:14:18 AM
#9
51% of the accidents will not happen, I believe bitcoin team.

Too late Tongue
newbie
Activity: 56
Merit: 0
June 14, 2014, 02:39:01 AM
#8
51% of the accidents will not happen, I believe bitcoin team.
copper member
Activity: 1498
Merit: 1499
No I dont escrow anymore.
June 08, 2014, 01:32:05 PM
#7
-snip-
Thanks to both of you for your explanations.

The damage that can be done with a >50% attack is pretty big. It's a lot worse than I had originally assumed. I'm surprised it's not more of an issue in the community.

Keep in mind that this would hurt the overall bitcoin value. So it would also hurt the attacker(s). It is an issue within the community. People are allready worried that ghash.io might accumulate over 50% hashingpower. They are currently @ 39% according to blockchain.info https://blockchain.info/pools

Also keep in mind that the biggest pools are in fact pools. There are many many people hashing for ghash.io. According to their own info1 55% of their hashingpower is from individuals. They could easily switch the pool info in their ASICs and work for a smaller pool.

[1] https://ghash.io/ghashio_press_release.pdf
legendary
Activity: 826
Merit: 1000
amarha
June 08, 2014, 06:55:52 AM
#6
Would just waiting for a larger number of confirmations defeat the attempts at double spending?

Like instead of the standard 6, instead wait for 8, 10, 12? Whatever the safe amount would be so that there would be enough blocks solved by different independent sources. Or am I missing something here and that wouldn't be a solution to the problem...

As has already been pointed out, someone with more than 50% of the total bitcoin network hashing power (in other words, more hashing power than the entire combined rest of the bitcoin network), then they have a better than 50% chance of solving the next block.

If they fail to find the next block fast enough, then they can ignore the block that the rest of the network is working and, and continue on their own block.  They still have a better than 50% chance of solving their block before the network moves on to yet another. Because their probability is better, sooner or later, they will solve enough blocks in a row to create a chain longer than the one they are trying to replace.  The higher above 50% they have, the faster they'll be able to get such a chain of blocks.  Once their chain is longer, their chain replaces the existing one.  Any transactions that were confirmed in any of the replaced blocks that aren't also in the attacker's blocks immediately become unconfirmed.  If the attacker spends any outputs in his blocks that were spent elsewhere in the chain being replaced, then those transactions in the replaced blocks don't just become unconfirmed, they also become invalid and they disappear.

They can start their attack chain many blocks before paying you.  Therefore, no matter how many confirmations you want to wait for, they can make sure that they have a long enough chain to broadcast once you've accepted the payment.

Thanks to both of you for your explanations.

The damage that can be done with a >50% attack is pretty big. It's a lot worse than I had originally assumed. I'm surprised it's not more of an issue in the community.
legendary
Activity: 3388
Merit: 4615
June 07, 2014, 09:01:03 PM
#5
Would just waiting for a larger number of confirmations defeat the attempts at double spending?

Like instead of the standard 6, instead wait for 8, 10, 12? Whatever the safe amount would be so that there would be enough blocks solved by different independent sources. Or am I missing something here and that wouldn't be a solution to the problem...

As has already been pointed out, someone with more than 50% of the total bitcoin network hashing power (in other words, more hashing power than the entire combined rest of the bitcoin network), then they have a better than 50% chance of solving the next block.

If they fail to find the next block fast enough, then they can ignore the block that the rest of the network is working and, and continue on their own block.  They still have a better than 50% chance of solving their block before the network moves on to yet another. Because their probability is better, sooner or later, they will solve enough blocks in a row to create a chain longer than the one they are trying to replace.  The higher above 50% they have, the faster they'll be able to get such a chain of blocks.  Once their chain is longer, their chain replaces the existing one.  Any transactions that were confirmed in any of the replaced blocks that aren't also in the attacker's blocks immediately become unconfirmed.  If the attacker spends any outputs in his blocks that were spent elsewhere in the chain being replaced, then those transactions in the replaced blocks don't just become unconfirmed, they also become invalid and they disappear.

They can start their attack chain many blocks before paying you.  Therefore, no matter how many confirmations you want to wait for, they can make sure that they have a long enough chain to broadcast once you've accepted the payment.
copper member
Activity: 1498
Merit: 1499
No I dont escrow anymore.
June 07, 2014, 10:33:38 AM
#4
-snip-
I thought there was an issue of probability involved with the reversing of transactions.

Thats true because there is probability involved when finding a block.

Or is it something like when the attacker who has >50% chooses to publish their blockchain that doesn't have the transaction that's supposed to be double spent the network automatically accepts their chain because they've done the most proof of work?

"The rule" is: the longest blockchain is the valid one. The attackers can work on an an alternative blockchain in secrecy and release it to the public when its long enough.

Does that sound right at all? So it doesn't matter how many confirmations happen because they will all be wiped out when the network accepts this new chain?

That is correct. The new, now longer chain of blocks, will most likely have no or very few transactions at all* thus all transactions that have been included in other blocks are now invalid. Well not exactly invalid, they never happened.

*this depends on the prefernce of the attckers, they can basically put any valid transaction in those blocks they want
legendary
Activity: 826
Merit: 1000
amarha
June 07, 2014, 10:14:35 AM
#3
More blocks = harder to replace them.

That will not help against a 51% attack (or >50% attack) because the idea behind that attack is that the attackers control the majority of the hashing power. They could create replacement blocks without publishing them and wait 8, 10, 12 or even more confirmations before broadcasting them.

I thought there was an issue of probability involved with the reversing of transactions. Or is it something like when the attacker who has >50% chooses to publish their blockchain that doesn't have the transaction that's supposed to be double spent the network automatically accepts their chain because they've done the most proof of work? Does that sound right at all? So it doesn't matter how many confirmations happen because they will all be wiped out when the network accepts this new chain?

copper member
Activity: 1498
Merit: 1499
No I dont escrow anymore.
June 07, 2014, 09:50:47 AM
#2
More blocks = harder to replace them.

That will not help against a 51% attack (or >50% attack) because the idea behind that attack is that the attackers control the majority of the hashing power. They could create replacement blocks without publishing them and wait 8, 10, 12 or even more confirmations before broadcasting them.
legendary
Activity: 826
Merit: 1000
amarha
June 07, 2014, 09:47:06 AM
#1
Would just waiting for a larger number of confirmations defeat the attempts at double spending?

Like instead of the standard 6, instead wait for 8, 10, 12? Whatever the safe amount would be so that there would be enough blocks solved by different independent sources. Or am I missing something here and that wouldn't be a solution to the problem...
Jump to: