New thoughts on Ethereum upgrade process, L1 expansion and data availability

Author: Dankrad Feist, Ethereum Foundation developer Source: X, @dankrad Translation: Shan Oppa, Golden Finance

1. Ethereum’s current upgrade process is problematic

Developers need a long-term commitment to network capacity, otherwise they cannot build on Ethereum. They need to be clear about what their plans are for the next five years. So I hope we can finally agree on a timeline within the core development process.

2. We wrongly claimed that Ethereum L1 cannot scale and L2 is the only solution

This view is completely "ivory tower" thinking - Vitalik published relevant content 3 years ago, and with the development of data availability sampling (DAS) and zkEVM technology, these elements are now in place.

While certain limitations do exist, it is entirely feasible to scale Ethereum 100-1000x within a few years. This enables many features that are not possible based on native L1 or other alternative rollups. I think really pushing this forward and stopping looking for excuses not to do it is the key to Ethereum's victory.

3. Not all transactions can be placed on Ethereum L1

Whether the actual upper limit of L1 is 1,000 tps, 100,000 tps, or even 1 million tps, it cannot meet all needs. Therefore, we still need to significantly increase data availability (Data Availability) and create a high-security Ethereum ecological safe zone for rollup. These rollups are allies of Ethereum and enable trustless use of Ether and other Ethereum assets.

But it all has to happen on a truly meaningful scale. We need to get to the current goal of 1 megabyte per second (1MB/s) in 1-2 years, not 5 years and then move on. I'm currently working on how to maximize decentralized network Data Availability Sampling (DAS), but until that's done, adding data availability providing nodes with higher requirements would be a good solution with security assumptions similar to an optimistic rollup (a Honest nodes) are the same.

Other data availability protocols are currently being developed to support GB/s systems! Ethereum doesn’t have to reach this level, but just reaching 1/1000 is not enough at the moment.

Some projects have publicly stated that they will never use Ethereum's Data Availability (DA), and have publicly advertised support from Vitalik (which is ironic). Accepting this status quo while still claiming that these projects are “Ethereum-aligned” lacks ambition.