Seeing Ghosts: Vitalik Is Finally Formalizing Ethereum’s Casper Upgrade

ghosts-halloween-e1504735178119
7 September 2017

Casper is starting to look less like a ghost.

While that might sound too good to be true for the long-waiting ethereum faithful, CoinDesk has confirmed that the network’s creator, Vitalik Buterin, is now in the process of crafting three white papers explaining Casper, the protocol’s much-anticipated version of proof-of-stake consensus.

As such, the papers could come to mark a major milestone for ethereum in that, while Casper has long been proposed as a better and greener way to keep the global network in agreement about the blockchain’s transaction history, the industry has been waiting for details to be put on paper.

Rather, the idea has been trapped in the brains of a few ethereum developers, with breakthroughs and development scattered across online chat groups and blog posts over the last several years.

As proof-of-stake is pitched as such a crucial piece of ethereum, users have had to trust these developers do indeed have a good plan.

As such, Buterin’s formal white papers will now subject ethereum developers to peer-review, which could mark a big step forward for the project which is currently getting ready to upgrade to Metropolis before changing the system over to proof-of-stake.

“In short, the Casper designs are getting better each iteration,” ethereum developer Virgil Griffith, a well-known hacker and ethereum developer who is currently reviewing Buterin’s white papers, told CoinDesk.

The papers have been hiding in documents in the ethereum research GitHub for the last couple of weeks, with Buterin and Griffith making updates every now and then.

However, it’s worth noting the early-stage nature of the work. The documents, which have question marks and some “coming soon!” bubbles scattered throughout, won’t be ready for at least a month, Griffith said, adding:

“I’m working on them as I type this.”

Slow and steady

The first paper, “Casper the Friendly Finality Gadget” – a play on the 1990s film “Casper the Friendly Ghost” – explains how the consensus system works at a high level.

Notably, the paper reiterates a recent change in direction for ethereum. Rather than switch from proof-of-work to proof-of-stake right away, the plan is start off slow by first weaving the two together.

The paper explains:

“The proposal mechanism will initially be the existing ethereum proof-of-work chain, making the first version of Casper a hybrid PoW/PoS algorithm that relies on proof of work for liveness but not safety, but in future versions the proposal mechanism can be substituted with something else.”

Expanding further, proof-of-work will be used to validate most ethereum blocks, but proof-of-stake will be used as a “checkpoint” for every 100th block, providing more “finality” to the system, or a guarantee that transactions cannot be spent more than once.

The paper goes on to cover possible attacks, such as “long-range attacks,” that validators could try to use and how Casper aims to overcome them.

Devil in the details

The other two papers go into the minutiae of the system.

Titled “Incentives in Casper the Friendly Finality Gadget,” the second paper explores the incentives that go into making the system work and ensuring it doesn’t get messed up.

For Casper, there are two types of things that can go wrong: “Safety faults” occur when the rules are broken, for instance when two validators come up with incompatible states. “Liveness faults” occur when the system stops, or fails to push transactions through.

Because users have to deposit some of their money to participate as validators, the paper outlines that if validators try to go against the rules, the system will steal their deposit.

The third paper, “Automated Censorship Attack Rejection,” focuses on 51-percent attacks – those in which a miner or mining pool accumulates a majority of the network computing power, and can then twist the rules of the system to their advantage to, say, double spend or block transactions.

Developers have thought a lot about this problem, since if it were to occur, it could reduce confidence in the blockchain as a single source of legitimate transactions.

In the paper, Buterin argues Casper mitigates this problem, since the protocol punishes attackers by taking away their deposits if they do something wrong.

Another approach

It’s worth noting, though, that Buterin’s approach to proof-of-stake as outlined in these white papers, isn’t the only approach.

The Ethereum Foundation’s Vlad Zamfir, who leads the actual development of Casper, said he plans to release more details about his version of Casper before Devcon, ethereum’s big developer conference in the fall.

He told CoinDesk:

“Vitalik is more driven to implement something soon, whereas I’m more driven to search for theoretically optimal solutions even if it means some delays.”

And again, because Buterin’s version is just now getting squared away on paper, it could be assumed that during the peer-review process, things will need to be further fleshed out.

Yet, developers have already started implementing the first step of Buterin’s Casper approach, which is slated to go live sometime after ethereum’s next big upgrade, Metropolis launches this fall.

And with that, it’s unclear whether the theory behind Casper will be fully ironed out before people see the ghost in real-life.

Ghost image via Shutterstock