Blockworks Research: Where will Ethereum go next?

Author: Donovan Choy, Blockworks Research; Compiler: 0xxz@黄金财经

Today, the price of BTC is $68,000 and the price of ETH is $2,700. The last time Bitcoin reached this price point was in November 2021, when the price of ETH was around $4,300.

At the time, Ethereum's expensive gas fee averaged 130 gwei (27 gwei today). This prompted the public at the time to announce "abandonment" of Ethereum and switch to other L1 chains-remember "FOAN" (Note: Fantom-FTM, Harmony-ONE, Cosmos-ATOM and Near-NEAR)?

You can easily count all the active rollups at the time. According to L2Beat, there are at least 105 rollups today, which cumulatively process about 296 transactions per second.

Ethereum L1 is doing ~13 TPS, so ~21x better than it was three years ago. That’s progress in every sense of the word.

This progress is largely due to the proliferation of Rollups through Rollup-as-a-Service (RaaS) providers like Caldera and Conduit, which make it very easy and cheap to create a Rollup.

The second major factor is the rise of alternative data availability solutions like Celestia, and the EIP-4844 Dencun upgrade in March 2024. The latter provides a cheaper way for L2s rollups to publish their batched data to L1 — ~10x cost savings.

All is going according to plan. But there’s still a long way to go to reach Ethereum’s ultimate goal of ~100k TPS on Ethereum mainnet and all L2s in the “Surge” phase of the roadmap.

There is still a lot of work to do to get there.

First, Ethereum needs PeerDAS (Peer Data Availability Sampling) to scale data availability. PeerDAS borrows some important ideas from BitTorrent’s peer-to-peer file sharing to handle more data by distributing fragmented pieces of data among Ethereum validator nodes without storing everything.

Then there is data compression to make the millions of everyday signatures for transactions we sign through our MetaMask wallets more compact so that the data usage of an ERC-20 transfer does not exceed 30 bytes (compared to 180 bytes today).

This can be achieved by using the BLS (Boneh-Lynn-Shacham) digital signature scheme to aggregate multiple signatures into a single signature that can be verified with multiple public keys at the same time, which is not currently possible with current ECDSA (Elliptic Curve Digital Signature Algorithm) signatures.

Next is the use of Plasma, a scaling solution where blocks are published off-chain, while the Merkle roots of those blocks are placed on-chain — a huge computational improvement over today’s Rollups, where full blocks are committed on-chain or via validiums.

And that’s just for L2.As for L1, execution needs to continue to scale, either by increasing the gas limit for L1 blocks or introducing piecewise upgrades that make the EVM more gas-efficient. Methods include tweaking the EVM object format (EOF) bytecode, or creating parallel compute, data, and storage markets (i.e. multi-dimensional gas pricing).

All of this ensures that L1 doesn’t become a “ghost chain” used only for settlement and DA. In Vitalik’s words, Ethereum still “needs to be robust enough to actually handle the highly complex and messy shutdowns of L2 at least occasionally,” which is why the “long-term security of the network” is so important.

Finally, there are UX improvements, like solving cross-chain interoperability issues so that using Ethereum doesn’t feel like using 100 different chains. This work is happening at the protocol level, through the development of better ERC standards like ERC-7683, and by having dozens of other players working in areas like “shared ordering” and “chain abstraction.”

All of these roadmap features must be done while maintaining decentralization, which requires pushing L2 toward trustless active proof systems — and keeping a close eye on centralization at the MEV layer.