Scroll, an open-source tech project working on making Ethereum more useful, is expected to go live in on the Ethereum mainnet this summer, says its co-founder, Sandy Peng.
Peng, a road warrior nominally based in Hong Kong, is part of a small, competitive group of companies and projects looking to create the standards on which Ethereum can scale – and become the rails for a growing range of financial activities.
Scroll is in a race to create a zero-knowledge Ethereum Virtual Machine, along with big players such as Polygon and Matter Labs.
Layer 1s and 2s
What’s a zkEVM? Let’s take a step back.
Ethereum is a “Layer 1”, meaning it is the base of a blockchain network that maintains transaction records. It’s where trades settle. A “Layer 2” is the application itself, such as a DeFi protocol or a trading algorithm: these are the businesses that use Ethereum to do stuff.
The “Ethereum Virtual Machine” is a grandiose name that reflects its ambition to be a universally accessible layer 1 for all kinds of businesses. (As opposed to Bitcoin, whose Layer-1 wiring doesn’t allow for simple add-on applications or uses.)
In practical terms, EVM does two things. First, is the protocol for smart contracts to be used on the Ethereum blockchain. Second, it then recomputes the state of the entire Ethereum ledger every time one of those smart contracts executes (ie, if there’s a trade).
Ethereum’s problems
This has led to a lot of problems, because EVM’s software is written in a computer language, Solidity, that was designed just to create Ethereum. But applications are built on other languages that are well known, easy to write, and offer more flexibility. So there’s one problem, which is that apps don’t translate readily to Ethereum.
The second, bigger problem is that Ethereum, because it is a public and permissionless blockchain (anyone can write on it or read it), requires all nodes to validate every new block and every transaction. This is very inefficient, and has meant processing trades is very slow and very expensive. This is considered one of the biggest impediments to Ethereum realizing its potential to serve as next-gen universal infrastructure.
The developer community is working on a laborious path toward addressing this issue. Last year’s Merge, in which Ethereum changed its consensus mechanism from Proof of Work (like Bitcoin) to Proof of Stake was part of that.
But more steps remain. One of these is sharding, in which pieces of a transaction can be randomly distributed to participating nodes and then recomposed, so nodes only have to validate a fraction of ongoing transactions.
Security and ZKs
This raises a new concern: security. In the world of blockchain, the technology faces a Trilemma: the impossibility of a system being decentralized, fast, and secure at the same time. By shifting to Proof of Stake (in which influence in validating blocks reflects a node’s possession of ether, ie, the block with the most moolah has more votes), Ethereum has exposed itself to conflicts of interest. Future efforts such as sharding also create new opportunities for transaction information to leak.
In comes the zero-knowledge proof, a bit of cryptographic theory in which a party can prove its information (such as a trade in all its details) without revealing what that information is. A ZK proof should allow provers to simply state that they possess information in order to be verified by others.
A zkEVM is therefore applying ZK cryptography to applications built on Ethereum, so that smart contracts can prove or verify information in a way that maintains a user’s secrecy.
This falls into a generic category of blockchains called rollups. These are blockchains that sit above a Layer 1 such as Ethereum, and also host apps. They process transactions in those apps (such as a DeFi trade) in a more efficient way – using various computing techniques to bundle trades or reduce node responsibilities – and then return the transaction to the Layer 1 for final settlement.
The ZK technology, which is just one type of rollup, has been put into practice by a few cryptocurrencies designed for complete privacy, such as Zcash and Monero. There has also been academic research around outsourcing this processing out of the realm of software, and shift it to hardware with chips optimized for such processes.
“If we could do the same thing for software, we could scale Ethereum by ten times,” Peng said. Two years ago she launched Scroll with the collaboration of the Ethereum Foundation (a non-profit association representing the developer community) and other projects, including Zcash.
Rolling out the rollup
Scroll has been active on its testnet (a sort of sandbox for blockchain projects) since August, testing what decentralized apps (or Dapps) can port to EVM without a hitch.
“We’re now processing 1 million transactions a day,” Peng said. “It takes only three seconds to produce a block. But we plan to increase our throughput by another order of magnitude before we hit the mainnet in June.”
The mainnet – live for the market – requires a lot of de-bugging. “We’re going through the governance with DeFi projects,” Peng said, noting there are now 900,000 unique wallets that have tested Scroll’s zkEVM.
“There’s a flurry of creativity,” she said. “There’s a new breed of Dapps with real-world utility.”
These Dapps focus on DeFi, gaming, and NFTs. One of the goals of using a Layer-2 protocol like Scroll is to make them compatible.
Efficiency versus security
The challenge Scroll faces as a business is competition: interoperability can work only among Dapps using a given zkEVM. There are other rollups out there. Polygon says it will launch its version this month. Matter Labs is hot on its heels. Might this leave Scroll in the dust?
The most likely winner will be Layer 2s that prove they are the most secure (as opposed to the ones that do the best job of marketing themselves). This remains a work in progress.
A major point of vulnerability is the “bridge”, the point at which transactions or assets move between Layer 1s and Layer 2s. A blockchain may be secure but a communication to another chain provides an opportunity for a hacker to steal keys.
Another vulnerability is the hardware, the physical circuits. As with other forms of computing, bugs can lead hackers close to the metal – exploiting bugs that let them listen directly to the traffic in a server.
“Our design is fully decentralized at every level,” Peng said. “We’re the first proven network” that is so decentralized that validators can operate Scroll at home on a GPU (a PC equipped with graphics processor chips). “They can package ten, twenty, a hundred transactions onto one block [on Scroll’s blockchain] and post it to the Ethereum block.”
She says Scroll is on track but won’t go to the mainnet until the team is satisfied with the protocol’s security. Scroll, like Matter Labs, is relying on open source development, so that it can be updated by developers from anywhere.
“We have a lot of third-party eyes on our robustness,” she said.
The trust factor
This leads to a third challenge for rollups: auditability. Being in the business of trustless verification, they still need to be verified themselves – and for now that requires an actual auditor, not just an open community. Last year’s crypto collapses have thrust the trust factor front and center.
Scroll’s auditing process is being run by an internal team. Whether it has the confidence to rely on open-source auditing remains to be seen.
But Peng says major DeFi protocols are now using Scroll to make their transactions more efficient, including Uniswap, the leading DeFi platform by volumes.
“We’re more confident,” she said.
Scroll’s business model is similar to Ethereum’s, to capture block auctions and charge fees for transactions. It is paid-for infrastructure, and does not issue its own token. Scroll last raised money in a $30 million Series A funding round in 2022 from Bain, Polychain, and other investors, both institutional and individual.
“I can’t comment on our funding plans but we have a decent runway until we find the right product-market fit,” Peng said. “We’ll be cash-sustainable in five to 10 years’ time.”
