2026-06-08 Logos storage weekly

Week 23 2026 (Mon June 01 - Fri June 05, 2026)

After identifying a critical vulnerability with our initial hidden service specification, an second iteration is now ready for review.
libp2p 2.0.0 has been released, and we are closely collaborating with Anon. Comms in getting it updating Mix to support it.
We gave a talk during this week’s ZK research call about “dynamic” proofs; i.e., proofs in which a large part of the witness remains unchanged across proofs, and techniques for speeding those up. Writeup available here.

achieved:
- Just keeping Draft PR feat: logos storage for History Archives status-go/pull/7486 aligned (rebasing and fixing rebase problems). Did not have manage to finish my own small review of the PR one before the end of the week.
next:
- Open for review.

achieved:
- Started working on a Mix simulation, both to clarify things and for the simulation results to feed back into the design
- Discussion on how to handle two Mix operational modes: ForwardAddr (exit != destination) and MixNode (exit == destination) (see https://discord.com/channels/973324189794697286/1511629352880574576). To support the discussion, some examples/proposal has been made:
  - Draft PR for experimentation logos-co/nim-libp2p-mix/pull/11 where destination read behavior is encoded in Sphinx packet payload and the libp2p_mix_experimental_exit_is_dest is removed.
  - Example for the ping protocol in MixNode mode of operation added: examples/mix_ping_mix_node.nim
  - Draft PR for the spec changes created (logos-lips/pull/351), but put on hold till the above discussion settles on something.
- Investigating/reviewing updating the current MIX implementation to libp2p v2.0.0 (in draft release currently):
  - logos-co/nim-libp2p-mix/pull/12 - in review
  - logos-co/nim-libp2p-mix/pull/14 - merged
  - logos-co/nim-libp2p-mix/pull/15 - merged
next:
- Finish and experiment with the simulator
- Implement connection state management in the reference impl.
- review logos-co/nim-libp2p-mix/pull/12
- continue discussion on Mix operation modes
- workout example of the simple Mix transport (keep postponing it, because something new continuously pops out with higher priority - like updating Mix to support libp2p v2.0).

achieved:
- Found (and hopefully fixed) an attack against hidden services with the current SURB-based design.
- Worked on second iteration on the hidden services specification:
  - Define hidden services epoch and update the descriptor and KDF used to match the implementation.
  - Changed the SURBs-based client-service communication and defined an approach that uses one-time rendezvous point since the previous approach turn out to be unsafe.
next:
- Studying the design choices and possible alternatives to (the various components of) the Tor design.

achieved:
- Support for --no-bootstrap-node flag in the storage node, allowing initial bootstrap nodes to startup without connecting to a default network. PR
- Fixed an issue where PVCs were being orphaned after release test runs PR
- Added dist test framework tests (meta tests) that cover the release test fixtures.
next:
- Finish initial dist test framework tests
- Address PR feedback (mainly, try to eliminate PVCs in GKE setup)

achieved:
- Gave a talk about “dynamic” proofs at the ZK research call. You can also read the detailed writeup.
next:
- We should implement and benchmark this approach, but it’s low-priority right now.

next:
- Track the upstream nim-libp2p-mix 2.x PR so we can merge anonymous DHT queries into master once master bumps to libp2p 2.x

achieved:
- merged Rahul’s multitasking rework into nim-groth16 (Nimbos runs on refc, for reasons…)

achieved:
- Multiple fixes in libplum accepted and available in release 0.6.0
- Provide several fixes in nim-libplum
- Use observed address to dial nodes in Autonat V2 in libp2p
- Update nimbus build system
- Update libp2p version to 2.x.x
- Update NAT Traversal PR and tests
next:
- Test on deployed nodes