Design a wide RPC engine rehaul
Design a wide RPC engine rehaul
This is a sub-milestone of %(2024Q2) - Layer 1 - Public RPC endpoint supporting average 1k RPS
The current RPC implementation, that relies on Cohttp and Resto, is know not to fit our needs (lacks of expressivity) and to suffers from several flaws (resource leaks). It is needed to well understand the issues of the current implementation to propose a alternative solution. After specifying the needs and flaws, we will propose alternative implementation for the RPC engine. Thanks to toy examples and POC, we will demonstrate the pros and cons of the alternative solutions to design a roadmap for a potential future project
Work breakdown
-
(ETA: 2024-05-03) Analyze the RPC API flaws induced by the current RPC implementation -
Requires: %Improve and consolidate RPC UX
-
-
(ETA: 2024-05-10) Search for alternative implementations -
(ETA: 2024-05-24) Propose POC/toy examples to simulate our main RPC use case (with Resto) -
Toy example using Cohttp only -
Toy example using Httpaf only -
Introduce Resto in toy examples if needed
-
-
(ETA: 2024-06-21) Implement alternative solutions through the POC/toy example -
(ETA: 2024-06-21) Demonstrate the strengths of the proposed solution -
(ETA: 2024-06-28) Design a roadmap (for Q3) toward switching to a new RPC engine