Meeting Summary:
The team reviewed ongoing network experiments and testing progress, including work on the Fussaka network, private mempool development, and Devnet 3 deployment with a new Prism fork. They discussed host key releases for the Holski scheme and planned future testing arrangements, while also addressing sync tests and benchmark releases with a new consolidated genesis format. The conversation ended with discussions on block-level access control implementation, EIPs related to gas repricing and multi-dimensional metering, and the handling of Ethereum call gas limits.
Click to expand detailed summary
The team discussed ongoing experiments with the Fussaka network, including non-finality tests involving full nodes and super nodes, which demonstrated successful recovery times of about an hour and 20-30 minutes respectively. Bharath reported progress on the private mempool, mentioning an issue with the get payload API transition that needs fixing, and suggested using the existing spammer jobs for testing, which Parithosh agreed to set up. The team plans to monitor the spammer jobs and peer-to-peer data, with Bharath committing to spend more time on the private mempool once the fork transition issue is resolved.
Parithosh reported on the progress of Devnet 3, where a new Prism fork with runtime hooks has been deployed for testing. The network is performing well with 1,500-1,600 nodes, and a BPO test is scheduled to check the system’s ability to handle high loads. The team discussed the performance of the network, noting that the test was successful. Dustin is working on a problem, and the team is looking into the cause of the issue. The team also discussed the performance of the network, noting that the test was successful. Dustin is working on a problem, and the team is looking into the cause of the issue.
The team discussed host key releases for the Holski scheme, noting that while community validators can support ad hoc releases, there were concerns about disrupting the normal release flow. They agreed to continue testing with Definite 4 and planned a shadow fork after final releases, with Parithosh requesting any testing accommodations be communicated in advance due to the time required to set up networks. Pawan suggested exploring longer non-finality periods for certain attacks, and Parithosh offered to set up a smaller test network for this purpose after Definite 4 testing concludes.
Parithosh reported on sync tests, noting a fix was needed for Nethermind and that Raphael ordered additional machines for testing. He mentioned a new page for tracking historic sync times across clients. Mario shared updates on benchmark releases, mentioning a new consolidated genesis format that allows for faster and more efficient testing. They discussed plans to extend this format to consensus tests in the future.
The team discussed implementation details for block-level access control and related testing. Toni confirmed that the execution specs exist and Teretto has started implementation in Git, but noted that a test environment is still needed. The team decided to activate block-level access control at the time of Glamsterdam and Toni will provide the necessary configuration. Parithosh suggested rebasing on top of Fusaka, but the team agreed to focus on getting a working devnet first.
The team discussed several EIPs and their implementation status, with a focus on gas repricing and multi-dimensional metering. Ansgar highlighted that the upcoming Glamsterdam fork will significantly impact network throughput and throughput levels. The team agreed to prioritize testing at higher throughput levels and prepare for Glamsterdam fork, which will significantly impact Ethereum’s throughput and throughput. The team also discussed the need for a shared development environment to test and validate the Glamsterdam fork’s impact on Ethereum’s throughput and throughput.
The meeting focused on discussing the implementation of gas limits for Ethereum calls, particularly in relation to EIP-7 and its impact on clients. It was agreed that the current wording of the EIP should remain as is, and the team will investigate the issue where the bug is located. The team agreed to leave the wording of the EIP as is and the decision was made to patch the bug.
Next Steps:
- Bharath to fix the fork transition issue and spend more time on private mempool testing.
- Parithosh to set up spamming on the private mempool and analyze peer-to-peer data to ensure get tools are properly disabled.
- Client teams to review and approve Tanno’s EIP for merging this week.
- Mario to make another benchmark release and share it with the CK team for testing.
- Toni and Parithosh to coordinate on creating a Kurtosis config for block level access list implementation.
- Justin to merge the intervals per slot PR soon.
- Ansgar to create a meta EIP for gas repricing topics related to Amsterdam.
- Mario to check with the team about the merged fix for the modexp base and mode 0 issue.
- Client teams to double-check coverage of gas cost calculation issues this week.
- EF Protocol Security team to continue reviewing clients for potential issues.
- Client teams to manually test Eth calls against all client RPCs to identify which client has mistakenly configured the tx gas limit for Eth calls.
Recording Access: