Zero-knowledge rising

Zero-knowledge rollups have been rising on the horizon for ages. Finally, last week, the arrival of a general purpose zk-rollup to scale Ethereum became more than a stop in a roadmap, with the almost simultaneous launch of zkSync Era and Polygon's zkEVM.

Let's go through the basics first.

Ethereum has a scalability problem with no quick solution in sight. If you still need to understand why, check out this post on the Scalability Trilemma.

Some time ago, Ethereum developers decided that the best way to tackle scalability problems was by relying on rollups to do the heavy lifting. Rollups are secondary blockchains capable of processing transactions, bundling them, and then sending the bundle to Ethereum for final validation. This turns a lot of transactions on a rollup into one single operation on Ethereum, saving time and money.

Rollups differ in many aspects, but one of the most important ones is how they perform validation. How do they guarantee that the transactions they've processed and bundled are valid so they don't send any contaminated bundle to the Ethereum blockchain? Two main trends emerged:

• Optimistic rollups are precisely that: optimistic. They give transactions the benefit of the doubt and consider them valid by default. But they rely on fraud proofs to ensure the computations done off-chain are correct. Before sending them for final settlement on Ethereum, they leave time for third parties to search for fraudulent transactions and report them (for a bounty). Optimistic rollups need to offer a buffer time for these external agents to operate, which makes the final validation slower.
• On the other hand, zero-knowledge rollups use cryptographic proofs to ensure that the computations done off-chain are correct without relying on fraud proofs. Zero-knowledge proof is a mathematical concept that allows a process to be proven valid without revealing information about the process. For example, imagine being able to answer the question "do you speak Chinese" with "yes," and that answer is enough. Math-gic. Short answers respect privacy and save a lot of time, which are very valuable in crypto.

An excellent example of zero-knowledge proof is this. How can I prove that I know where Waldo is without actually telling you where Waldo is? Using a big sheet of paper with a hole in it, I can confirm that I know where he is, but without telling you how I got there.

EVM compatibility is another critical feature that tells the difference between alternative takes on scalability.

The Ethereum Virtual Machine, or EVM, is a virtual software stack that runs in Ethereum's nodes. The EVM is similar to server software, installed in the decentralized network of Ethereum nodes instead of in physical servers. The EVM gives the Ethereum blockchain its programmability because it allows the execution of transactions and smart contract code.

The ability to interact as frictionlessly as possible with Ethereum's Virtual Machine is an outstanding feature of any Layer 2 solution. Being fully EVM compatible means that 100% of the code that can run on Ethereum will run on a rollup. This creates benefits in terms of network effects and developer complexity. But there are more efficient machines than the EVM. As a result, some rollups might tweak the code to increase performance to the detriment of compatibility.

EVM compatibility comes in many flavors. Below is a chart created by Vitalik Buterin to explain that the different types of zero-knowledge rollups were on the horizon according to the tradeoffs made between performance and EVM compatibility.

Zero-knowledge rollups bore the promise of a scalable, secure, safe, and cheap Ethereum. But the technology was taking ages. Some minor manifestations of zero-knowledge chains existed in app-specific blockchains, but a general-purpose zk-rollup was only in our dreams. Matter Labs had zkSync running since December 2020 but with marginal use. Starkware had Starknet, a general-purpose zkEVM.

Then we had many announcements: Matter Labs had a general-purpose zk rollup in the works, and so did Polygon, home of one of the most popular Ethereum scalability solutions (Polygon PoS). Later, Scroll and Consensys joined.

But the announcements ended, time went by, and we all got back to our business, forgetting about zk rollups and getting excited in turn with other rollups (did you guys get ARB?)

When suddenly, two of the announcements became flesh last week. Two and a half, we might say:

• Matter Labs launched zkSync Era public alpha (renamed their old zkSync, zkSync Lite).
• Polygon launched their zkEVM "beta mainnet": open to users but to be handled with care.
• Consensys joined the party by announcing a dev-ready test of Linea, their stab at a zk rollup.

Here's what we know about each one.

zkSync Era was open to devs since October 2022 and moved to public alpha in March. The blockchain has been thoroughly audited and has already attracted $25M TVL (it was $19M last time I checked, and the number will probably have changed a bit when this post goes live). Let us leave you with just the curve, then.

Era will still be mainly centralized in its first steps, but the team is committed to decentralization, so we will likely see a token coming up.

zkSync Era has some exciting features coming up, like account abstraction and a significant UX improvement in crypto that will, among many other things, allow users to pay transaction fees in the token they choose.

Matter Labs announced that more than 200 projects, including Chainlink, SushiSwap, Uniswap, Aave, Argent, 1inch, Gnosis, and Curve, are already committed to deploying on Era.

Polygon's zkEVM also launched last week, with Vitalik playing a central role in the ceremony—quite an endorsement.

Polygon's rollup is moving slower than Era. Probably because Polygon speakers asked users to proceed with caution, given the state of the development. TVL is still super slim ($250k when writing).

Decentralization has yet to be seen, but the team is committed to it. Initially, fees will be paid in ETH, and MATIC will eventually step in to participate in staking and validation.

But one of Polygon's zkEVM's most promising features is its EVM compatibility: while zkSync Era is a Type 4 rollup, according to Vitalik's chart, Polygon's zkEVM is a type 2: compatibility is greatly enhanced, which will bring Ethereum's network effects.

Linea. To be fair, Consensys' zk-rollup does not belong in this announcement batch because they only launched a dev-ready testnet. But they deserve some credit for reaching more than 1.5M transactions after launching their first version of a testnet in January. Also, we appreciate them using an actual word instead of a salad of meaningless initials. Also, Consensys has Metamask and a history of excellent developer performance, so they'll bring something exciting.

In sum, it's mostly zkSync who's walking the talk. Era has the first-mover advantage in the zk-rollup race and has the difficult task of dealing with everyone's expectations regarding zero-knowledge scaling. However, they can also benefit from the secondhand experience of Arbitrum and Optimism. We wish them the best.