Summary

This first week was consacred to things I forget/didn’t known/not well understand about Ethereum.
I also started to think about the project to work on. I wanted a micro-economics one (RIG opened questions) but I think I would not have enough time for that kind of subjects (research topics).

Ressources

• RLP serialization: already know about that, but I needed a reminded.
• RLPx transport protocol: I didn't there was a dedicated protocol for node communicatio. Honestly I am not an expert of transport protocols so it like a pain-in-the-*** to read. I found a python implementation that is not much clearer.
• ephemery.pk910: one of the project is an ephemery testnet; and I remember that for my job, we needed to find a testnet that give users a lot of ETH; an intern find this project, so I re-took a look at it. It centralized (1 person manage the node) but a good starting point. It seems to be the current implementation of the ephemery-testnet from Ethereum.
• Engine API: after Mario presentation about Ethereum, I discovered a new service I didn't know: the engine API which make the CL to communicate with the EL; like to get payloads, block hashes etc.
• Engine API geth implementation: that's the go code for the interface design here. It refer to the EL here and may be called by the CL. I also understood how it lives along side with the execution API. I was wondering if the node.Node could facilitate the deployment of a REST API wrapper of the JSON-RPC API but it requires a rpc.API type, so it might not help. I found a TODO that may interest some people that might read this note .
• Deneb consensus-specs: I am monitoring the changes on the consensus-specs related to validators rewards/penalty. 2 changes may impact their income:
  • get_attestation_participation_flag_indices: there is no more inclusion delay for the target checkpoints; it means validtor can publish a very late attestation for the target checkpoint and still have a reward (before they had a penalty).
  • process_attestation: the inclusion range for an attestation is between the current and the previous epoch.
    => validators that publish attestation at the beginning of an epoch will have more time for attestation publication before being penalized; but since validators attest for each epoch, it's not unfair.
    I don't think it's directly link to the Deneb update: EIP7045. I don't know what are the security consern on the LMD-GHOST so I asked to Danny Ryan.
• EIP7045: see above
• LMD GHOST + bitwise LMD GHOST: I did need a huge reminder of GHOST on Ethereum (it's been a long time since I dove into consensus algorithms). I did that to be sure to undestand why the consensus-specs were cahnged regarding the target checkpoints. LMD GHOST is a fork choise rules based on the subtree that have the latest message. The bitwise version seems to be quite complicated, but this python implementation make it more understandable. (note for myself: What is called the virtual height is the number of blocks to make the link between the mother and child through the bitwise decomposition).

Questions

nil

TODO

• Find an unused RaspberryPie 8gb and setup a new Ethereum node (will 8gb be enough for testing purpose ?). Maybe I will my company node or Infura IDK.