Ethereum at 100,000 TPS: StarkWare Discusses the Future of Layer 2 Scaling

StarkWare co-founders Eli Ben-Sasson and Uri Kolodny sit down with Crypto Briefing to discuss how their project’s STARK-based technology will reduce the cost of using the Ethereum network.  

What Is StarkWare?

StarkWare is one of several crypto projects that leverages Zero-Knowledge Rollups to help scale Ethereum. Unlike other Layer 2 scaling solutions, it uses ZK-STARK proofs, otherwise known as zero-knowledge, scalable, transparent arguments of knowledge. StarkWare’s President Eli Ben-Sasson and Chief Architect Michael Riabzev co-invented STARKs. 

StarkWare’s has an application-specific scaling solution called StarkEx that uses STARKs to achieve scalability. StarkEx powers several major platforms including the decentralized exchanges dYdX and DeversiFi. Over the past six months, it has settled over $250 billion worth of transactions and with significant gas fee savings for users. 

Starkware is currently developing a multi-app Ethereum Layer 2 solution called StarkNet. The product will let users hold their funds in a single wallet and interact with multiple applications on Layer 2. Additionally, StarkWare intends to launch the current StarkEx platforms as so-called “Layer 3s” on top of StarkNet, making it easier and cheaper for users to interact with these applications. 

STARK-based technology offers several key innovations over existing Ethereum scaling solutions. STARKs offer a way to use Ethereum permissionlessly at a high speed and low cost. STARKs are also the first cryptographic technology that allow proofs to be verified without any form of trusted setup. 

Crypto Briefing caught up with StarkWare co-founders Eli Ben-Sasson and Uri Kolodny to hear about their plans to scale Ethereum on Layer 2, and they discussed how the technology offers security for users because it is completely trustless. “The math that we developed prevents anyone from stealing funds—it’s impossible,” Ben-Sasson said. “We can’t steal funds from our customers even if we wanted to, or even if we were hacked.”

“We can give Darth Vader or the Bogeyman access to our servers, and there’s still no harm that they could inflict on users’ ownership and custody of their assets,” Kolodny added.

StarkWare’s goal is to unlock blockchain technology’s true potential. While cryptocurrencies have attracted more mainstream than ever in recent months, blockchains like Ethereum are constrained by high costs and limited data availability. As a result, most newer users are currently priced out of the network. Other Layer 1 chains increase transaction throughput by running fewer yet more powerful validators. However, this method increases centralization as users need to trust a small set of validators for all transactions on the network. StarkWare is aiming to stay true to Ethereum’s founding principles of decentralization and accessibility through its scaling solutions.

Ethereum mainnet currently handles about 15 transactions per second, and more complex smart contract interactions can cost hundreds of dollars at peak congestion. StarkWare says that StarkNet will reduce gas fees by a factor of 100 to 200, while Ethereum co-founder Vitalik Buterin has previously stated that rollups like StarkNet will help the network achieve 100,000 transactions per second. 

Staying True to Ethereum

StarkWare is currently responsible for proving and sequencing all transactions on StarkEx and StarkNet that are sent to Ethereum for confirmation. While this has some benefits, such as limiting the potential for MEV, it doesn’t fulfill the company’s vision of decentralization. 

By the end of 2022, StarkWare plans to make the sequencing and proving software open to the public, allowing anybody to participate and secure transactions made on StarkNet. Ben-Sasson shared his optimism about the shift to community validation, stating:

“I think it will likely look very much like mining on Ethereum in the early days. You’ll have to get off your couch, learn a little bit, install some things, and run a server. But hopefully, you won’t need to build a facility in Iceland next to some geothermal plant or something. It’s going to be within reach.”

Like many other successful crypto projects, StarkWare will need to have an incentive structure in place to get people engaged with proving and sequencing on StarkNet. Similar to how miners receive block rewards and transaction fees for validating transactions, StarkNet will also incorporate similar monetary incentives. “There will be fees that will go to, among others, the operators, provers, and sequencers, and then beyond that, we’re deliberating on other approaches,” says Ben-Sasson. “It will be a fully decentralized network that will require coordination and governance mechanisms,” added Kolodny. 

Crypto projects often issue tokens to achieve decentralization and add governance mechanisms. In 2021, several popular Ethereum projects such as dYdX, Ethereum Name Service, and ParaSwap launched their own tokens with airdrops for early users. Many Ethereum users have speculated that Layer 2 projects such as StarkWare will also issue tokens to encourage adoption, but Ben-Sasson and Kolodny did not shed any light on whether StarkWare was planning to launch one.

One of the key reasons for Ethereum’s success has been its commitment to decentralization and accessibility. Anyone with a few graphics cards can start mining blocks and validating transactions, and even running a node requires simple hardware. By emulating Ethereum’s accessibility, StarkWare is also refusing to compromise on decentralization as it works toward its vision of a secure and public scaling solution.  

Security and Cost

StarkWare’s STARKs can be deployed in one of two data availability modes: Rollup or Validium. The differences between these two modes highlight the compromise between security and cost in Ethereum scaling. 

In Rollup mode, every transaction or change of state on Layer 2 is “rolled up” together and sent to Ethereum mainnet in a validity proof, meaning that they all benefit from Ethereum’s security. As this method uses more data and thus more block space than Validium, it costs more gas.

On the other hand, Validium does not report every change in the Layer 2 data to Ethereum. Instead, it relies on a data availability committee to confirm they all have the same state, before signing off the transactions along with the Merkle root of the new state. The small compromise in security results in substantial gas savings compared to Rollup mode.

Currently, applications such as dYdX run STARKs in Rollup mode to take advantage of its enhanced security. As the dYdX exchange handles billions of dollars in trading volume, it makes sense to pay for extra security. While Rollup mode currently reduces transaction costs by a factor of 100 to 200, savings that will increase as more people use the network due to gas cost amortization, there is still a linear limit to how much they can scale. It’s with Validium, though, that the true potential of Ethereum scaling can be unleashed. 

Additionally, when more transactions take place between the proof updates to mainnet, the gas cost gets split between more users. In other words, the more users Validium attracts, the cheaper the transaction fees become. 

The NFT-based soccer game Sorare already runs STARKs in Validium mode to finalize transactions. As games are likely to be some of the most gas-intensive protocols on blockchains over the next few years, Validium’s limitless scaling offer huge promise for blockchain gaming. Kolodny said that some game developers have already started to see the potential of the technology. He explained: 

“Gamers are telling us, in a very explicit fashion, ‘for the first time as a developer on the blockchain, I can actually focus on the game I want to build, as opposed to the computational resources I’m constrained by, or by the “gas ceiling” that’s hanging over my head. The question of ‘what was that game we wanted to build on Ethereum’ is not the relevant question; the real question is ‘what was that game that we actually dreamed of.””

StarkWare and Ethereum’s Scalable Future

StarkWare’s application-specific scaling solutions are already proving that STARK-based Layer 2 solutions could be a game-changer for Ethereum and the crypto space at large. StarkEx applications are already settling over six million transactions weekly, while the total value of transactions settled has surpassed $300 billion—orders of magnitude higher than other Ethereum Layer 2s such as the Optimistic Rollup solutions Optimism and Arbitrum. 

Ethereum’s roadmap is geared toward building out Layer 2 to help the network scale (on the completion of Ethereum 2.0, it will also add 64 new chains called shards). Vitalik Buterin has long discussed how Ethereum of the future will leverage ZK-Rollups. Kolodny summed up StarkWare’s ambition to help the blockchain achieve scalability, commenting:

“What we’re doing with validity proofs is making the most succinct use of this public utility—the blockchain—that we are aware of, allowing others to use this resource in a far more effective fashion.”

The next step for StarkWare is to bring StarkEx’s scaling power to its multi-application network, StarkNet. Once launched, StarkNet will unlock more possibilities for scaling, including cross-Layer 2 bridges that immediately make funds available to end users. This should help pave the way for shared liquidity across multiple Ethereum Layer 2 instances, improving efficiency and the cost of capital further than any other Layer 2 scaling solution. 

As the crypto space has grown over the last few years, Ethereum has soared in popularity. However, it’s also faced many well-documented scalability challenges. Thanks to StarkWare, the leading smart contract network may finally have a shot at becoming a truly scalable base layer for Web3.

Disclosure: At the time of writing this feature, the author owned ETH, IMX, and several other cryptocurrencies.