Work Done

In the last two weeks I did a lot of coding in relation to two main projects:

• https://github.com/Keep-Reth-Strange/reth-crawler: we reached 19 stars and I’m already really proud of it, although it’s not stable right now. By the way, the improvements in the last two weeks are:

  • API server completed and integrated into the same binary of the crawler. It serves a standard RESTful API to be queried.
  • I refactored the code from two different crates (reth-crawler and api-server) to just one binary crate.
  • I abstracted the database so that it’s super easy to change the db if we ever need to do it. I created a PeerDB trait and also 3 implementations of it:
    • Amazon dynamo DB implementation. This is the default database we are using now.
    • SQLite DB implementation. There’s a --sql-db flag to use this instead of dynamodb.
    • In memory db implementation. Actually I think I’ll delete it since it’s not very useful. SQLite already creates a local db file and does not require any server or configuration.
    • 0xprames (anonymous contributor) did a great job in finding some issues related to the small amount of nodes we previously kept collecting. We were listening only to DiscoveryUpdate::Added events and now we added also DiscoveryUpdate::DiscoveredAtCapacity ones. He’s also trying to find some strategies to score peers in order to speed up the crawling.
    • 0xsmith (another anonymous contributor) is working on the frontend in order to display those data we collect.

• EVM bytecode analysis: this is a new project that Georgios from Paradigm suggested some time ago. The goal is to analyze EVM bytecode in Ethereum mainnet and find some recurrent tuples or triplets or quadruplets of opcodes and maybe in order to be able to super optimize the EVM execution, maybe optimistically pre-loading some instructions and stuff like that. This is very similar to what happened with CPUs over the last decades…

  By the way the first issue was creating a command in Reth that analyzes bytecode and print opcodes frequencies. The PR is almost ready and I’ll push it in a couple of days max.

  I really like this projecy but the goal is very ambitious and the time for EPF is ending. So the plan is trying to do what is possible until Devconnect and then continue the work after EPF ends.

• https://github.com/alessandromazza98/evm-from-scratch: As a side I started this project of creating the EVM from scratch. I forked a famous repo and I’m doing a Rust implementation of the EVM. Super excited. I’ll write a daily update of it.

Near Term Goal

• reth-crawler
  • add new info for every peer added, such as:
    • synced / not synced
  • find some scoring strategies to optimize the crawler
  • refactor / fix bugs
  • help for the frontend
• EVM bytecode analysis: push the first PR and asks Georgios (or Rakita) what they suggest for next steps.