Topic: 2013-11-05 BBC.co.uk: Bitcoin at risk of network attack, say researchers (Read 1496 times)

No, this is an incentive to get all pools to strategise their Bitcoin client connections amongst other clients that are the best connected, best located and have the lowest latency network hardware. An incentive that already exists. That all the pool operators are more than aware of. Just because someone can write "She has the ability to detect new blocks on the network extremely effectively because she has controls a lot of nodes with low-latency, high-bandwidth connections" in a sentence, it doesn't magically become possible to create it at will. Even if "Selfish" Miner does their absolute best to develop their LowLatencyDreamNet^tm, every other pool operator is working to do the same, within the same constraints.

So basically a big pool like BTCguild might be able to pull it for a few hours or so until everyone realized they were submitting multiple blocks together and moved to other pools?

Found very good explanation of this. This is about the way of getting more miners to your "selfish" pool.

Define "success".

Sure, you can do the attack successfully, but with less than 51% hashrate, you'll burn more good block solutions than the number of successful attacks. Stupidest "attack" you can come up with, you end up attacking your own mining profits, just so you can "cheat" (and end up cheating yourself).

Perhaps trying to hack pools and gain access to their miners and intercept communication with pooled hardware would be the easiest way to obtain 51% of the network.

This can be mitigated by many spread out pools with distributed hash power; however, I'm guessing again that this wouldn't be easy, but perhaps just as "easy" as any of these other impossible feats?

For it to be a consummate success, you need 51% of the network as well, not just 25+ or 33+.

The likelihood of solving 2 blocks consecutive blocks (and holding the first back from propagation while you're working on the next) is very low for someone with less than 50% of the network hashrate. Much less likely than some other miner solving the current block from underneath you, nullifying all the work you've done on your potential cheaters block (it will not be valid as the solution to the honest miners block). If the "selfish" miner tried this multiple times, hoping they might get a lucky streak, they'd be statistically bound to end up throwing away alot of good block solution rewards, all just for the sake of getting a better chance of solving consecutive blocks? The probability of creating a self replenishing chain of blocks becomes smaller with each new consecutive block, < 50% hashrate simply cannot be that lucky.

So if you're a miner with malign intent, I think there are more things to worry about if you end up running > 51% of the network. There are much more "profitable" things you can do, but again, it's self-defeating; you'd quickly become 100% of a network that no-one was using, and of a cryptocurrency with a substantially reduced exchange rate. 51% is only any good as a takedown, not as a profitable move. The idea that you'd use 51% for the Cornell attack is beside the point.

Tl;dr: Selfish to the point of stupidity, such that you gain nothing, isn't actually selfish at all. Cornell selfish miner attack can only be a failed attempt at selfishness. 51% makes it a success. Then that fails. Fail no matter what, basically.

It's not a 51% attack, but you do need a sizable fraction of the hash rate in order to be successful.  According to the algorithm the longest block chain wins.  No need to publish it immediately.

1.    Find a successful hash but do not publish it.

2.    Start working on the next block.  While your pool is working on the next block, everybody else is still wasting time working on the previous block.

3.     Find the next hash and publish (or not and try for three ...).  Even if your second hash comes after someone else finds another hash to the previous block, your block chain is a block longer than theirs.  You win.  If you're ahead two blocks (which would not be rare if you have 25+% of the hash rate) then you have even more time to game the system.  

It's basically doubling down.  You won't win all the time, or perhaps not even most of the time, but  if you have 25%+ of the hash capacity you increase your odds greater than your share of the total hash capacity on average.

right. you'd need a pool or else secretly have hundreds of millions of mining hardware.

and for u to keep hashing power you'd have to assume your users at the pool didn't find out what you were doing

For this to be successful, wouldn't you need an extraordinary combination of short-sighted malice from the Bad Guys combined with utter obliviousness from everyone else not involved in the attack?

this was addressed in satoshi's white paper

The incentive may help encourage nodes to stay honest. If a greedy attacker is able to assemble more CPU power than all the honest nodes, he would have to choose between using it to defraud people by stealing back his payments, or using it to generate new coins. He ought to find it more profitable to play by the rules, such rules that favour him with more new coins than everyone else combined, than to undermine the system and the validity of his own wealth.

http://libertyblitzkrieg.com/2013/10/31/five-years-ago-today-bitcoin-was-born-read-satoshis-original-white-paper/

paper is here

http://arxiv.org/pdf/1311.0243v2.pdf

the only issue is that they assume that selfish miners do not care about what selfish mining would do to the value of bitcoin... sign...

let me summarize for anyone:

beginning of article: an attack on the network could happen from a vulnerability exploited by selfish large miners

end of article: the attack is very unlikely and almost improbable

I think it's yet another typical Bankster Backed Corporation (BBC) bullshit slur story.

Ignore it. 

I think what they are implying is that a big pool could try and solve a second block without telling anyone that they already solved a block, but that is basically a big gamble that they can solve a second block before someone else solves the first block and announces, which is not likely unless they have 51%?

It's basically a gamble, and it would be obvious if a pool started regularly submitting two blocks close to each other.

I think this IS a 51% attack in disguise, and the scenario it describes is based on a faulty assumption.

Not only does this attack require at least 25% of the network to even start it, the attackers need to somehow coopt or trick additional 26% of miners to join them in the attack. It is all based on the assumption that the miners are "selfish" to the point of happily joining the attack for a minuscule short term benefit regardless of any long-term consequences. Such as potential destruction of the very network they are deriving their income from.

So, in reality this should play out like this. One of the major pools (such as BTCGuild or Eligius) suddenly starts acting weird, announces an intention to screw everyone and take over the network. Not only that, they invite everyone else to join their Evil Empire asap. "Selfish miners", out of their natural greed and malice, abandon the "honest" network and join up in droves until the "evil network" reaches 51%, overpowers honest miners and starts wrecking chaos...

How realistic is this scenario? Would not it be more realistic for the miners to move from the "rogue" pool instead, depriving it of any hashpower it previously held?

Why is anyone even discussing it seriously as something "novel" is beyond me...

Just found they're discussing it here:
https://bitcointalksearch.org/topic/majority-is-not-enough-bitcoin-mining-is-vulnerable-324413

Not sure about the attack, doesn't make sense to me how it would work... If you didn't report blocks found then no one else would acknowledge your "find"? Basically, they would be fake?

Can someone clarify this?

http://www.bbc.co.uk/news/technology-24818975