Didn’t think I would post an update a mere 2 weeks after the initial post: https://www.reddit.com/r/ethereum/comments/mjyvxz/ethereum_scalability_race_april_2021_update/. I highly recommend reading that post for a primer on
But with the ETHGlobal Scaling Ethereum summit currently ongoing, we have tons of news on rollups. It took a few days, but I finally managed to get through the key presentations. Definitely worth watching if you have a cool 7 hours to spare: https://www.youtube.com/watch?v=Hsksm_by7LQ
Or, here are some highlights:
Matter Labs announced details about zkPorter. Essentially, this is a Validium-like solution – a ZK rollup architecture but data availability is off-chain. Where it improves on Validium is that the data availability committee is partially decentralized with its own proof-of-stake consensus mechanism. They haven’t released details about the consensus mechanism yet, but we know there’s slashing mechanisms involved and it’s implied to be real proof-of-stake, not delegated-type like pretty much every non-ethereum L1. zkSync 2.0 uses a Volition-like architecture where developers and/or users can seamlessly choose between Rollup or zkPorter modes for their desired level of security vs fees trade-off.
While you do compromise on security with zkPorter, it still offers far superior scalability, decentralization and security characteristics than any alternate non-ethereum L1. The worst case scenario is a super-majority 67% of zkPorter validators collude to freeze your funds – but it also freezes their own stake.
I have one strong criticism for Matter Labs, though – their marketing here is tone-deaf, by dubiously antagonizing optimistic rollups. Like I mentioned above, what they should really do is aggressively market against alternate L1s. There’s currently $250 billion+ misallocated into many of these ghost chains, and almost all of them don’t hold a candle to a programmable ZK rollup in terms of scalability, security or decentralization, let alone all three. By adding zkPorter to the mix, we now have scalability that’s completely impossible with L1s unless you centralize your validation to supercomputers like Solana is doing – at this point, it’s no longer a decentralized, trustless network.
Of course, Matter Labs is not alone – StarkWare is building a similar solution for StarkNet, where they will have a data availability consensus mechanism to add to Validium. However, StarkNet is only planning to get there in 2022, while zkSync 2.0 will potentially have the entire stack live on mainnet by late 2021. There’s plenty of exciting progress for StarkNet before then, though. In May, we will have the StarkNet single-app rollups – what they call Planets – live on testnet. Once Planets are done, anyone will be able to release their own arbitrary code using the Cairo language, on their own single-app rollup. Where it gets interesting is Constellations, where it brings a multi-app rollup this is due to release in late 2021.
On the point of Validium data availability consensus mechanisms, it’s important that data sharding (what was previously “Phase 1” and now due after The Merge) will bring massive data availability to L1. However, I can still see zkPorter/Validium being relevant to potentially accelerate low-value high-throughput applications like gaming or social networks (which are currently impossible using any blockchain tech) to the tune of millions of TPS.
Optimism and Arbitrum both have their public release-candidate testnets out now. Optimism is taking a more conservative approach with guardrails and whitelists for contract deployment on mainnet. A common misconception is that “Optimism is delayed till at least July”. In reality, they have been live on mainnet since January. As per their whitelist approach, we’ll see Synthetix to gradually move more of their functionality over to L2 over the next couple of months, and I expect to see Uniswap V3, Maker Protocol and a handful of others join them there. The July timeline is for when Optimism turns off the whitelists and anyone will be able to deploy smart contracts. It’s also important to note that this whitelist is for developers only – users can freely use Optimistic Ethereum. Meanwhile, Arbitrum’s key advantages over Optimism are they have no such whitelists and their testnets and upcoming mainnet deployment will be open to all from day one. Arbitrum also supports BLS signature aggregation natively already – the same cryptographic innovation that makes >120,000 validators possible on beacon chain today can also be used to dramatically improve batching and scalability. Of course, Optimism has the industry support at this time, but it’ll be curious to see this space develop.
Optimism’s approach of aligning its clients with Geth should mean it’s the most Ethereum-like environment of the bunch. They fully plan to implement L1 innovations like EIP-1559, statelessness to further enhance Optimistic Ethereum. Interestingly, “zkEVM” is in the works for Optimism. Though Karl (whose enthusiasm is forever infectious) doesn’t go into details there – it is pretty clear the Optimism team is aware of the benefits of ZK rollups and fully plan to leverage the tech in the future in some form.
Finally, Polygon is also working on both optimistic and ZK rollups, and other non-rollup solutions, though their roadmap remains pretty vague for now.
Summing up: zkSync 2.0 may be stealing the show, but they need to prove their tech on testnet – which happens in May. We’ll get a clearer picture then. Of course, Optimism is a force to be reckon with, and likely where most blue-chip DeFi projects will deploy first over the next 3 months or so. It’s the closest to Ethereum L1, and for complex smart contracts that wish to maintain as close to like-for-like parallel deployments on L1 and L2 as Uniswap and Synthetix do, it’s going to be the preferred option. Arbitrum could potentially be the first publicly available programmable rollup, and is still very closely aligned with L1 with the sweet benefits of BLS signature aggregation. StarkNet’s tech is already proven with DeversiFi, dYdX and ImmutableX live on mainnet, and their methodical approach may prove pragmatic long-term. While zkSync 2.0 and StarkNet may diverge from L1 EVM more than Optimism, it also opens up potential benefits of L2-specific efficiency and scalability. I can totally see deployments for things like non-financial applications that skip L1 entirely and live purely on a ZK rollup and perhaps even on zkPorter/Validium. Indeed – this is what ImmutableX is doing.
Overall, the Scalability Race continues to heat up. It is still early, and each of these players have their own sets of strengths and weaknesses.
Despite all the noise around Dogecoin, BSC etc – this is the real narrative that is currently unfolding in the blockchain space that’s going to have lasting impact on the industry for years to come. On that note, it’s very important to note that the rollups journey is just beginning, and it’ll be a space that’ll take several years to mature. Make no mistake, though, this is the present and future of blockchain scalability.