The Ethereum Shanghai upgrade may be postponed to April, and EIP4844 has been renamed to Deneb

Original Title: 《Ethereum All Core Developers Consensus Call #102 Writeup》
Author: Christine Kim, galaxy
Compiled by: Overnight Porridge, DeFi Dao

On February 9, 2023, Ethereum developers gathered for the 102nd All Core Developers Consensus (ACDC) call.

Note: The ACDC call is hosted by Danny Ryan from the Ethereum Foundation. It is a bi-weekly series where Ethereum developers discuss and coordinate changes to the Ethereum protocol. The ACDC call primarily focuses on the development of the Ethereum consensus layer, while another series of calls (the All Core Developers Execution (ACDE) call) focuses on coordinating code changes to the Ethereum execution layer.

This week, the consensus layer (CL) client teams discussed the testing progress of the Shanghai and Capella upgrades. Recently, the Shanghai upgrade testing progressed smoothly on the Zhejiang testnet, and developers agreed to initiate upgrade testing on the next public testnet, Sepolia, on February 28.

Before the official Ethereum mainnet upgrade, Sepolia is the second-to-last testnet. Based on Sepolia's activation date and discussions regarding additional testing for MEV-Boost, the mainnet launch date for the Shanghai upgrade seems more likely to occur in early April.

Goodbye Zhejiang Testnet, Hello Sepolia Testnet

On February 7 (Tuesday), the Shanghai upgrade was first activated on the public Ethereum testnet named Zhejiang. Barnabus Busa, a DevOps engineer at the Ethereum Foundation, stated that this activation was "very successful" and emphasized that the network encountered no issues during the upgrade process. Marius van der Wijden, a developer of the Geth execution layer (EL) client, stressed that all execution layer client teams should create dedicated tests to ensure the correct encoding and decoding process for staking ETH withdrawals on the network.

ACDC call chair Danny Ryan asked the group if they felt there was a need for more specialized testing on withdrawal credential changes. Based on the group's silence, it seemed that the execution layer client teams were satisfied with the testing work they had completed. Ryan continued to discuss specialized testing for MEV relays and builders. Barnabus Busa mentioned that he had reached out to the Flashbots team, hoping they would assist in testing MEV-related infrastructure and software, but had not yet received a response from Flashbots. Terence Tsao from the Prysm (CL) client team noted that the lack of "capable relays and builders" on the testnet was a major obstacle.

Due to insufficient testing of the MEV-Boost software, fallback mechanisms when the MEV Booster software fails, and updates to the specifications for builder and relay operators, some developers in the call expressed reluctance to test the Shanghai upgrade on more public testnets until these specific testing areas were addressed. A developer from the Prysm (CL) client, using the pseudonym "Potuz," stated, "I think we shouldn't even be talking about forking Sepolia without first testing the builders." In response, ACDE call chair Tim Beiko countered that Ethereum core developers have historically not allowed the readiness of infrastructure providers to affect the timing and testing of protocol layer upgrades.

After further discussion among developers, Danny Ryan suggested that the execution layer client teams focus on testing the MEV software components that directly impact the Ethereum protocol layer. Additionally, Ryan agreed to collaborate with other members of the Ethereum Foundation testing team to work on a dedicated test network to assess validators' responses to MEV infrastructure and software failures during the Shanghai upgrade.

Ryan stated:

"My intuition is that we won't stop testing on Sepolia right now, but rather parallelize many of these testing efforts and continue evaluations over the coming weeks."

It is reported that Sepolia is the second-to-last public testnet. Based on the readiness of the client teams, developers agreed to schedule the activation of the Sepolia Shanghai upgrade for February 28. Tim Beiko emphasized that this means a blog post about the Sepolia upgrade will be published on the Ethereum official website around February 20, and all client teams should have a working version of the Shanghai upgrade ready by next Friday (February 17) at the latest.

Assuming the activation progress of the Sepolia testnet goes smoothly, developers will schedule the activation date for the Shanghai upgrade on the Goerli testnet, which may be released sometime in mid to late March, while the Ethereum mainnet upgrade is expected to occur sometime in early April. Ryan mentioned that Ethereum core developers could simultaneously decide on the activation dates for the Goerli testnet and Ethereum mainnet after the activation of the Shanghai upgrade on the Sepolia testnet.

Goodbye EIP 4844, Hello Deneb

While the Shanghai upgrade testing work is ongoing, developers are preparing to activate Ethereum Improvement Proposal (EIP) 4844 in the next upgrade following the Shanghai upgrade. As noted in previous call records, the Shanghai upgrade is the name of the upgrade that enables staking ETH withdrawals on the Ethereum execution client (EL), while Capella is the name of the Ethereum consensus client (CL) upgrade. In the last ACDC call, developers agreed to name the EIP 4844 upgrade Deneb (Tianjin Four Star), which is the name of a first-magnitude star in the constellation Cygnus. Hsiao-Wei Wang from the Ethereum Foundation reminded the consensus client (CL) team that starting this week, the naming of EIP 4844 on the relevant GitHub version will be updated to Deneb, so developers who have opened pull requests under the EIP 4844 name should update as soon as possible.

Anton Nashatyrev, the chief developer of the Ethereum blockchain explorer ether.camp, presented his findings on decoupling block production from blob production during this week's call. Here‌ summarizes Anton Nashatyrev's findings, and the previous week's developer call summary provided background on the discussion regarding the separation of blocks and blobs. One of the main benefits of using separate sub-networks to propagate blobs and block data is that these messages can be communicated more efficiently across the network. Anton confirmed that when decoupling these messages, the message reception time was reduced by 40% - 50%. Developers discussed ways to improve these simulations and preliminary investigations around blob validation strategies. An anonymous developer from the Lighthouse (CL) client team, "realbigsean," also emphasized a public issue‌ regarding the beacon chain API and endpoints for signing blob transactions. Developers participating in the call agreed to read his issue and provide feedback.

Other Topics

Other miscellaneous topics raised during this week's call included:

Call for volunteers to become EIP editors: Tim Beiko encouraged execution client (CL) developers to join the EIP editing team to help review EIPs related to the execution client (CL).

Call to deprecate outdated engine API calls: Mikhail Kalinin, chief researcher at ConsenSys R&D, proposed several different approaches to deprecate the outdated engine API call known as "exchangetransitionconfiguration." Given that this code change is not urgent, developers agreed to revisit the topic after the Shanghai upgrade is completed.

SSZ call: In recent developer calls, discussions around SSZ format changes have become increasingly relevant to the preparations for Deneb. Therefore, developers will organize a dedicated call on February 15 to discuss matters related to SSZ format changes.