pragma-org/amaru

Buidler Fest Workshop Feedback - Feature request

Dam-CZ opened this issue · 1 comments

Dam-CZ commented

MoM of the Workshop organised on the 23.04.2024, transform (if accepted) these items into other issues

Pain point 1 : Observability and operability of a Node on Cardano
Pain point 2 : Lack of customization of the content and structure of a Node

  • Request 1 : Make it easier to run a node (less demanding of infrastructure & resources)

  • Request 2 : Have an observer Node that just connects and starts storing data and blocks (something like Slurp made by Pi)

  • Request 3 : Be able to choose which Node you want to connect with (configuration)

  • Request 4 : gRPC & Mini Protocols integration (start with pulse check, light wallet...)

  • Request 5 : Implement Mythril to run the Node from a snapshot of the chain

  • Request 6 : Produce state of the chain as a proof for both the Rust Node and Mythril

  • Potential action : Gather feedback and benchmark infrastructure usage with the modular approach (Prioritize KPIs linked to RAM & Disc usage)

  • Potential action : Have workshops with the SPOs to gather direct feedback of usage and target improvements

  • Nice to have request : Trustless setup on mobile phone to run part of the Node

Dam-CZ commented

Features taken into account in the Amaru roadmap from the feedback of the Buidler fest :

Represent ledger data : Use a trusted ledger state to bootstrap from that point (option : Mythril)
Mini protocols integration : Build in the Node which allows the user able to chose which interface and modules to run (i.e. Rosetta as an option)
Archiver node : Have an observer Node that just connects and starts storing data and blocks (something like Slurp made by Pi)

Actions for high level targets:

Measure the cost of running a node (infrastructure and demands) and improve from the current Haskel nodes
SPOs workshops to gather feedback of resource usage and improvement targets for the Amaru Node

Already part of the roadmap but mentioned :

Light relay approach renamed into "Ouroboros proxy"
Not depending on a single upstream Node to trust
Modular design of the Node

Additional workstreams to be linked with the roadmap:

Workstream related to creating ZK proofs of events as they happen? (to be consolidated with Sebastian dcSpark)