Ethereum cohort week 16

Updates:

  • Last week i successfully wrapped up Kev's idea: https://hackmd.io/@6iQDuIePQjyYBqDChYw_jg/H1xXvMatq of having verkle crypto API, and posted in on separate repo: https://github.com/dragan2234/rust_verkle_crypto_api/ . So i can easy experiment with it, optimize it, and port the primitives to besu native repo: https://github.com/hyperledger/besu-native
  • I still want to fix few things, but I'm happy with the work. I think JNI can definitely be used for besu since rust is more performant then Java for cryptography. The only issue is that there's an overhead of 5-10ms when calling a rust function. Maybe this can be optimized even more.
  • Regarding verkle in-circuit, I was thinking if different field could be used for pedersen so it's even more SNARK/STARK friendly, but my conlcusion is that we should stick with bandersnatch since proof systems are improving so fast that there are even field agnostic proof systems and we are already commited to bandersnatch work. Very interesting work and comments: https://x.com/Charles_Chen533/status/1721359542599180406?s=20
Select a repo