Vac R&D Roadmaps and Reports

Search

SearchSearch

2025-04-21 Vac weekly

Apr 21, 2025, 6 min read

Vac 2025/04/21

vac:p2p:

  • ift:2025q1-gossipsub-perf-improvements
    • Submitted stagger send article on arxiv. The article is available online
    • Worked on large message handling (GossipSub v1.4) revision. Need 2 days to finalize.
    • Worked on GossipSub v1.4 specs feedback. Working on revised draft based on feedback
    • Spent some time on Nimbus integeration in Ethshadow. Now its terminating on unsupported system call signalfd4 (working on adding support for it)
  • vac:p2p:ift:2025q2-nimlibp2p-kad-dht:routing-and-rpc
  • vac:p2p:acz:2025q2-acz-consulting:libp2p-mix-consulting
  • vac:p2p:ift:2025q2-nimlibp2p-autotls:certificate-request-auth
    • Working on a standalone autoTLS script which contains:
      • Partial ACME client to request a challenge from Let’s Encrypt
      • PeerID auth
    • Opened a PR on nim-jwt to allow for flattened JSON JWT conversion, which we need for communicating with the ACME server
    • Some observations:
      • There are no nim ACME clients (had to build half of one), we could write one, but we’d have to maintain it too
      • Nim RSA key handling support is lacking a bit
      • We had to implement peerID auth
  • vac:p2p:ift:2025q2-nimlibp2p-webtransport:http3
    • finished wrapping nghttp3 lib -> nim-nghttp3 repo
    • ci, more precisly setting up nghttp3 source, is not working for windows. i’ll come back to this at later stages after webtransport is implemented. since debugging this took much time and is not really required to make progress on webtransport front (as unix based ci is working).
  • vac:p2p:ift:2025q2-test-coverage:qa-consulting
    • assisting radek in navigating improvments in nim-libp2p tests.
      • proposed improvments that are likly to half total test execution time.
        • ci completing in 15-20min and windows platfrom 23-25min, previusly >30min and >50min respectivly
      • most proposed improvements are still not even tested -> there is still a lot of room to improve total tests execution time (in local and ci)
  • vac:p2p:ift:2025q2-nimlibp2p-maintenance:maintenance
  • misc/admin
    • 1 CC 2d OOO
    • 1 CC 1d OOO

vac:tke:

vac:qa:

  • waku:2025q2-status-go-functional-testing:chat-actions-tests
  • waku:2025q2-interop-testing:jswaku-filter-hardening
    • Created new package and testing it in the CI
  • nomos:2025q2-da-testing:protocol-api-compatibility
  • nomos:2025q2-da-testing:interaction-data-flow
  • nomos:2025q2-e2e-testing:maintenance
  • vac:2025q2-nim-libp2p-testing:gossipsub-extended
  • vac:qa:waku:2025q2-waku-sync-testing:unit “25%”
    • Test environment in nwaku repo has some issues . COuldn’t run the new added tests
    • Try soltuions suggested by Simon
    • had to make new config.nim file with added dependicies and making changes to make the new tests run in PR
    • Added finger print tests but not able to execute them yet due to errors in the environment
  • admin/misc
    • Sick leave 6 CC days
    • Public holiday 2 CC days
    • PTO 1.5 CC days

vac:dst:

- 1 CC OOO Friday 18th Apr
- 1 CC OOO Monday 21st Apr
  • vac:dst:status:2025q2-status-go-evaluation:store-performance
    • Discovered other issues with light status-backend nodes
  • vac:dst:vac:2025q2-libp2p-evaluation:regression-testing
    • Help P2P debug issues.
    • Coordinated with p2p/acz team about node refactor refactor
  • vac:dst:ift:2025q1-vaclab:concurrent-experiments

vac:sc:

vac:acz:

  • ift:2025q2-zerokit:rln-execution-bottleneck

    • reviewed all related PR and small research regarding how Merkle tree was implemnted in zerokit

  • ift:2025q2-rln-status-l2:maintain

    • discussion with team about the starting tasks

  • ift:2025q2-libp2p-mix-testnet:dockerize

    • Fixed Mix instantiation issues [WiP] (dockerized-mix-gossipsub-dockerfile-1 branch)
      • Introduced separate paths for private/public node info
      • Implemented new Mix protocol init/load functions for proper peer discovery
      • Implement connecting to all GossipSub nodes (not just Mix nodes) [WiP]

  • ift:2025q2-libp2p-mix-testnet:update-rfc

    • Started RFC refactor [WiP] (mix-rev1 branch)
      • Re-architected Mix node roles to isolate core protocol from Entry/Exit layers
      • Clarified Mix protocol scope: message-based routing only
      • Rewrote Mix node roles to clearly define sender/intermediary/exit roles
      • Added pluggable component and mix core components section

  • nomos:2025q2-nomos-consulting:zk-consulting-nomos-1

  • ift:2025q2-zerokit:parallel-computation-study

    • Continued work on the parallelism feature for rln-wasm:
      • Finished benchmark results for rln-wasm parallelism report on browser (Comment)
      • Added browser tests and benchmarks for single-threaded and multi-threaded modes of rln-wasm (PR)
      • Encountered an issue with wasm-bindgen-rayon: the CPU maxed out and the program hung when running tests in multi-threaded mode. Reported it to the crate maintainer (Comment). This PR may be delayed until a response is received.

  • ift:2025q2-zerokit:merkle-tree-improvement

    • Start working on the task of refactoring the merkle tree using the rayon crate.

  • ift:2025q2-zerokit:iden3-witness-review

    • Got to a consensus of what to do next for the task
    • Open issue about perf regression
    • Open issue about crate publishing

  • ift:2025q2-zerokit:rln-execution-bottleneck

    • PR review & merged
    • A small PR for unit testing poseidon hash computation
    • Another small PR about documentation
    • Another PR with more optim for Poseidon hash
    • Discuss about a followup (// algo for OptimalMerkleTree)

  • ift:2025q2-rln-status-l2:maintain

    • Read about Linea architecture
    • Meeting “Status L2 <> ACZ RLN Deployment” + discussion about the starting tasks
    • Start a design for the RPC service modification + Prover + SC (unfinished, not sure where to put such very early document)
    • Having a discussion for the design

  • admin/misc

    • continue to work on the problem with incorrect verification PR
    • 1 cc is ooo for 4 days (day off)
    • 1 cc is ooo for 1 day (bank holiday)

vac:nim:

vac:rfc:

vac:sec:

vac:nes:

Graph View

Backlinks

  • No backlinks found