Ethereum Casper: First version now released. Here’s how it works
Security via economic theory. Get your head around Ethereum Casper and blockchain proof of stake.
The first version of Ethereum Casper FFG (Casper the friendly finality gadget) has now been released, largely in line with previously published specifications apart from a few clarifications.
Casper is a proof-of-stake version of Ethereum in contrast with the current proof-of-work version.
- Proof of stake (PoS). A system where the network is secured by funds held in online wallets. The wallets themselves use people's computing power to run the network, while the funds held in the wallet serve as a kind of "tangible weight" to make network attacks unrealistically expensive.
- Proof of work (PoW). A system, such as the one used by bitcoin, where the network is secured by computing power directly and miners compete with each other. The cost of computing power itself serves as the "tangible weight" in this system.
The main advantage of PoS compared to PoW is that it's a lot more cost and energy efficient. This is because it doesn't need to dedicate so much computing power towards serving as a "weight," and can instead just put it towards processing transactions.
The main disadvantage of PoS compared to PoW is that it's more dependent on economic theory for network security, which is trickier and riskier than "solid" computing power. As such, many of the Casper specifications describe a programmed economy to maintain system security.
"More than just the research team is using the contract now – auditors, client devs, etc – so we wanted to start issuing clearer versioning and changelogs to help everyone stay organized," said Ethereum developer Danny Ryan on Reddit.
Some of this clarification was focused on "slashing", which is part of the economic-theory security system. Slashing refers to the penalty for nodes that attempt to attack the network.
"If a slashing condition is found, the client creates and sends a transaction to slash on the Casper contract. The first transaction to include the slashing condition proof slashes the validator in question and sends a 4% finder's fee to the transaction sender," the release says.
- "Slashing conditions" are highly specific suspicious behaviours that indicate an attempted network attack. For example, if a validator deliberately misreports transactions.
- "Validators" are equivalent to miners on the bitcoin network. They hold funds in a wallet and use their computing power to process network transactions.
Basically, an important part of Ethereum Casper is a semi-automated bounty bot. Anyone can look for slashing conditions, and then report them to the bounty bot to automatically earn a 4% finder's fee.
The 4% finder's fee refers to 4% of the violator's confiscated stake. The stake is the money, the tangible weight, that validators have to put up as collateral if they want to help secure the network. The reason they're willing to do this is because they can earn about 5% interest per year on their stake as a validator. This interest comes from newly created Ether, just like bitcoin miners are in it for newly created bitcoin.
Initially, the minimum amount required to become a validator is 1,500 ETH, equivalent to about US$1.1 million at the time of writing. As such, violating any of the slashing conditions would be an extremely expensive mistake. Meanwhile, there's about US$40,000 going to the first person to spot and report an attempted violation.
The idea is that these twin motivators, the punishment for attempted attacks and the reward for spotting them, are enough to secure the network.
Will it work? Potential issues
Casper will initially be rolled out as a kind of second layer to the existing PoW Ethereum network. Essentially, it will use the current PoW system for a set of tried-and-tested training wheels while transactions are run through it to see if it performs as expected and that it is as secure as it should be.
One of the main doubts some have is that the slashing conditions aren't as comprehensive as believed, and that someone might find a way to attack the network without violating one. This is highly unlikely because blockchain attacks have been very thoroughly categorised and documented over time. Any attack that manages to poke a hole in the slashing conditions would almost certainly have to be brand new, entirely unheard of and outrageously creative.
Another doubt is whether the bounty system will be enough motivation for someone to actually keep an eye on the validators. There's a four-month window (the minimum amount of time one has to put up funds as staking collateral) to find and report them, but the question is whether people will bother.
After all, if slashing conditions are prohibitively expensive, it means no one will try to attack the network. But if no one tries to attack the network, then there will be no bounties available, and if there are no bounties, then there's no point in watching the validators, which opens the door to attacks. Plus, down the line, minimum staking requirements are expected to drop to about 32 ETH (US$25,000), so the chance of a 4% finder's fee (US$1,000) might not be as great an incentive.
These are also unlikely to be issues because it's relatively easy to set up a program that can automatically watch all the validators to check for slashing conditions, and automatically report any. Coupled with the four-month window, this seems relatively safe. Plus, a lower minimum staking requirement means a lot more validators, which means a much greater chance of discovering slashing conditions.
The main limitation might be human error. For example, if everyone downloads one popular "slashing violation checker" program that turns out to be compromised, or if everyone eventually turns off their slashing checker program because they never find any.
The Casper FFG system is largely designed to program human behaviour, but this is also what makes it potentially unpredictable.
Disclosure: At the time of writing, the author holds ETH, IOTA, ICX, VEN, XLM, BTC and NANO.