Scroll Aims to Be the Turtle That Wins the Ethereum Scaling Race

The most recent crypto boom pushed the Ethereum blockchain to its limits – with an unprecedented inflow of new users slowing down the network and pushing its traffic-based transaction fees sky high. This network congestion – and the accompanying fees and speeds – made Ethereum unworkable for some users, spurring a wide field of faster, cheaper blockchains bidding for Ethereum’s throne as the main hub for decentralized finance, non-fungible tokens (NFT) and blockchain research.

Ethereum’s core developers have long worked to scale up the network – refactoring its core code to boost speeds and decrease fees. But these interventions have only had a modest impact towards easing Ethereum’s expense and sluggishness, with more ambitious changes to Ethereum core code still years away. At this point, most of Etheruem’s core developers argue that third-party companion networks, called rollups, will need to be the primary method by which Ethereum maintains its status as the primary port of entry for the next billion crypto users. These rollups are separate blockchains that bundle up user transactions and write them in bulk to Ethereum’s transaction ledger.

But not all rollup networks are created equal. All rollups aim to “inherit” Ethereum’s security, meaning they all have special systems in place to (eventually) make transacting on them functionally equivalent to transacting on Ethereum itself. The most advanced breed of rollups, zkEVMs, have been competing to use cutting-edge zero-knowledge (ZK) cryptography to accomplish this goal.

Nearly a year into Ethereum’s “zero-knowledge” arms race, two teams that promised to deliver faster, cheaper crypto transactions using the complex tech say they’re gearing for (or at least nearing) some form of launch. For Polygon,“mainnet beta” is right around the corner. For Matter Labs, registration for its new “era” is already here.

Both companies are betting their respective versions of a zkEVM – a layer 2 blockchain that Ethereum-based projects can utilize without much effort – will herald the next generation of scalability for the world’s largest smart contracts platform. First to the finish line gets bragging rights and media glory – not to mention a whole host of users that come with a “first mover” advantage.

It’s a race that one other company called Scroll is more than happy to lose.

In an interview with CoinDesk, cofounders Sandy Peng and Ye Zhang repeatedly returned to the virtues of taking time (and plenty of it) to do the job right – as if the others were not. They fashioned Scroll as an open-source, open-door, community-driven entity that models itself in the shape of Ethereum itself. After all, the goal is to “help Ethereum scale” by using zkEVMs.

ZkEVMs are a kind of Ethereum rollup network – a layer 2 companion blockchain that records transactions to the base layer 1 Ethereum chain, but offers lower fees and faster speeds to users. In general, rollups work by bundling up big groups of transactions and then writing them to Ethereum in one fell swoop – drastically reducing the fees that would be required to make those transactions on Ethereum one at a time.

The bread and butter of any rollup network is its security system – the mechanism it uses to guarantee that the data it passes to Ethereum reflects real user activity.

The first rollups to market, such as those from Arbitrum and Optimism, operate using a dispute mechanism. When these networks pass a transaction down to Ethereum, watchdogs can dispute suspicious-looking activity within a set period of time (typically around seven days).

Upcoming zkEVMs, a more advanced breed of rollup, use zero-knowledge (ZK) proofs for their security. Rather than rely on third parties to call out bad behavior, these rollups use fancy cryptography to proactively prove that the transactions they pass down to Ethereum have not been tampered with.

ZkEVMs have been a long time coming. Early rollups that used ZK proofs were all application specific, meaning they were restricted to specific crypto use-cases (like sending tokens from one blockchain address to another). Newer zkEVMs such as Scroll aim for compatibility with any Ethereum app – meaning NFT marketplaces, decentralized lending platforms, and blockchain-based gaming apps should all be able to port their code over to these cheaper rollup networks.

Scroll is hardly alone in its zkEVM mission and it knows it will not be the first to get there. With big competitors putting varying stages of a “final touch” on their renditions of a zkEVM, the only thing certain about Scroll’s platform is it will not win any awards for development speed. At the moment, much of the general attention for zkEVMs are trained squarely on what Polygon and Matter Labs are bringing to the table.

Better to be right than fast

As CoinDesk has documented, not all of Polygon’s or Matter Labs’ claims map precisely with reality. Both companies brag that their zkEVM launches are imminent, but the platforms they are introducing to market in 2023 will come with severe technical handicaps. (For instance, both networks will rely, in part, on trusted intermediaries for their security – rather than pure zero-knowledge cryptography – when they go live).

It’s not lost on Scroll’s founders.

“We cannot ignore the marketing noise,” Zhang said during an interview. He said there will always be this or that claim to being first but Scroll is determined to “walk our own path” even if it comes at the expense of shipping fast.

The path they picked is slathered in molasses. Scroll is aiming to achieve a zkEVM that mimics the Ethereum Virtual Machine in every way possible. They want it so close to the real thing that builders on top of it won’t know – or even have to know – the difference.

Zhang said the team started building a zkEVM two years ago. Back then the consensus opinion on the feasibility of using ZK proofs to build a general purpose platform was just beginning to warm. It takes a lot of computing power to feed the complex cryptography underpinning ZK proofs; Zhang said generating a proof could take as long as a day. There was no guarantee of success.

“People still build only application-specific zk-rollups, like a DEX,” he said, referring to a decentralized exchange. “People think general purpose stuff is just so hard to build. And it’s very, very, very low efficiency.” But two years ago there were some breakthroughs in cryptography and also hardware acceleration, he added, that are now allowing the technology to zip ahead.

Scroll is taking a longer road by design. They are building “at the bytecode level” in order to “behave exactly like the Ethereum Virtual Machine” in every way, shape and form, according to Zhang. (Bytecode is the machine-readable form of code that dictates precisely how a program is to execute).

Being that granular can be one helluva drag on efficiency: Zhang says that the more equivalent one is to the EVM, the higher the “overhead” on the program. Higher overhead means more of a slog, and a more expensive slog, too. That’s prompted others in the zkEVM race to make efficiency “trade-offs,” by using compilers or other virtual machines. The result may be faster, but it’s not as compatible a design, he said.

To Zhang and the other employees at Scroll, the pain is worth the gain. They view the painstaking process of hammering out bytecode level equivalence as critical to their entire operation – it’s their raison d’etre.

“If you are bytecode level equivalent, you can be seamlessly compatible with all those toolings without any modification,” he said, referring to the Ethereum ecosystem’s “robust” developer toolkit landscape.

Simply put, a developer on top of a zkEVM won’t have to sweat the details or spend any time thinking about how their code will perform relative to how it does on Ethereum. For example, if you’ve got a relatively battle-tested protocol that’s been running for a few years on mainnet without a hitch, you can be sure it will do the same atop Scroll’s zkEVM. It will do exactly the same tasks and perform precisely in the same way. No extra audits or compilers necessary.

In crypto, what can go wrong will go wrong. It’s too soon to tell what might break in the nascent zkEVM space, but the founders of Scroll sounded sure that when it comes to their EVM-compatible-but-not-equivalent competitors, Murphy’s Law is destined to take the wheel.

“At the point of deployment, the difference is not visible. But where it will be majorly felt is when there is a security breach,” said Peng.

The lengths that Scroll is going to in its quest to replicate the Ethereum experience are borderline severe. Every detail of the tooling stack is reused, Zhang said.

Transparent, every step of the way

To make these matters harder (worse?) it’s all being done in the open. Every Scroll pull request, comment, line of code and so forth is playing out in real time on its Github repository. That means anyone with an internet connection – anywhere in the world – can track the progress of Scroll’s codebase. It’s a little flourish of transparency that Zhang said is “a step above” the commonplace practice (even among open-source projects) of building code behind closed doors and then throwing them open at some point in time.

The recipe is catching on. Since announcing its “alpha testnet” during EthDenver, over 550 “hacks and projects” have deployed, Zhang said. So far it’s all running quite well, even if – critically – it’s not yet a full representation of what Scroll aims to be.

Incomplete why? Because Scroll isn’t calling its testnet a “zk-rollup testnet” until the proofs are ready to rumble – and so far they are not. Zhang contrasted this “philosophical” decision with the more gung-ho approach seemingly employed by Scroll’s competitors in the zkEVM space. Without naming names, he said “some testnets don’t have a proof” even though their host projects insinuate the liveness.

The tension between Scroll’s meticulous mantra and its competitors elsewhere in this bustling corner of crypto nerdism was palpable. It is perhaps a product of the yin-yang lane Scroll has cast for itself. Even as it sticks to its tortoise philosophy – or at least to talking up the virtues of slow-and-steady in the press – its captains are deeply uncomfortable with the fast-and-loose hares.

When asked about how Polygon – the current leader poised to launch some zkEVM something before anyone else does – is framing its achievements in the press, Zhang said Scroll “has already pointed out some of Polygon’s over-claims [sic] from the marketing side.”

“They kind of oversimplify” when making claims about “100% compatibility,” he said, adding “we know how many tests they have ignored.”

The team is even uncomfortable with talking about talking about this dynamic, however. Before a reporter could follow up, Peng interjected with a cheery demand: “Let’s move on from this question.”