Project proposal: Embedding Reth into Grandine

[mjzk]

No description provided.

[sauliusgrigaitis]

Hey @mjzk, thanks for the proposal. I was thinking a lot about it and my understanding is that the first main step should be to get something working. This means that:

1. Inter-process integration via Engine API: A trivial approach that maintains separation between Grandine and Reth processes.

should be preferred over:

2. In-process integration: A more complex but potentially more performant approach that deeply integrates Reth into Grandine.

Once the trivial approach is confirmed we can talk about the more complex approach. I also suspect that the trivial approach could be good enough. The main motivation would be that in this approach it will be way easier to keep both Grandine and Reth code up to date. I'm also not sure which approach is better - integrate Reth in Grandine or integrate Grandine in Reth. However, if we have embeddable Grandine then we can embed it not only into Reth, but also into other EL clients. Maybe we can even have a call and discuss it further.

[taxmeifyoucan]

+1 to Saulius. I think the proposal is sound but also considering it's last month of the fellowship, the MVP with basic approach is more feasible. There was a similar proposal by @ThreeHrSleep who decided to drop it, he might provide some feedback https://github.com/eth-protocol-fellows/cohort-five/blob/7bd4509acaec8416926c5bc5f23c4b38e76c57b4/projects/direct-integration-of-lighthouse-reth-and-tracing-integration-in-lighthouse.md

[mjzk]

@sauliusgrigaitis @taxmeifyoucan Ok. No problem. Yes, the trivial way has its own advantages. I will change the proposal for focusing on the first integeration.