“The threat of a 51% attack still exists on proof-of-stake as it does on proof-of-work, but it's even riskier for the attackers. A attacker would need 51% of the staked ETH (about $15,000,000,000 USD).”
I'm pretty much just asking the same question on every thread now. How does something get slashed? These articles say nothing about it. There are many docs saying all you need is 1 honest slasher node. Well, what happens if you have a dishonest slasher node?
There is no separate notion of slasher, every staker is potentially a slasher. Slashing evidence can be included when you make a block. Doing so is rewarded, so you are economically incentivized to report slashable offenses if you find them. Blocks are validated and "attested" by everyone else. Failing to produce a block that everyone else accepts (edit: everyone else meaning at least 2/3 which is the quorum used) has at best the economic opportunity cost of what you could have earned and missed on and at worst (if someone else reports a slashable offense on your part that is accepted) a slashing penalty plus you get kicked out of the validators.
The rest is game theory. Dishonest validators win the game being dishonest if they are over 2/3 of the validator set. Then they can enforce their bad blocks into everyone else. Otherwise they get slashed or suffer other kind of penalties (inactivity leak if they manage to prevent the network from reaching consensus for a while).
So if you manage to bribe or buy 2/3 of the staked ETH then you can attack the network. There is of course a very significant economic cost to it. More interestingly at that point you are strongly aligned with the network itself as you represent a huge % of its value so you would be attacking yourself in a way. This is a notable difference with PoW where miners are not necessarily aligned in economic incentives with token-holders or users.
The network which was just shown to be 51% malicious actors? And how is it evaluated? What if the slasher is dishonest? What if it sends trillions of fake requests?
Evaluated in the same way that “does this address have enough to send to this much to this other address” is evaluated? If it isn’t valid, then a block containing it isn’t valid, and this is a thing which any client can automatically check.
What happens if you try to send trillions of invalid transactions in any sensible chain?
To whatever node your client is talking to, the invalid transactions should not be relayed. This is more of a DDOS situation than anything novelly technical. Like all computer systems, DDOS is a technical challenge that requires careful consideration.
Do you mean by claiming that the block containing it is invalid when it is valid, or visa versa? Or something else?
Also, how can there be a fee to trying to send an invalid transaction? Such a fee would have to be sent as part of a transaction, yes? Isn’t it just that someone spamming the network gets dropped from the gossip protocol stuff?
51% is possible, but if it happens someone has wasted (at current valuations) ~$15B gaining a majority stake in a network they are torpedoing. It would be pretty dumb to do something like that to your own investment.
https://en.wikipedia.org/wiki/Proof_of_stake
Start here.
“The threat of a 51% attack still exists on proof-of-stake as it does on proof-of-work, but it's even riskier for the attackers. A attacker would need 51% of the staked ETH (about $15,000,000,000 USD).”