Skip to content

003-exclude-rsocket-implementation.md

Latest commit

 

History

History
46 lines (34 loc) · 2.44 KB

003-exclude-rsocket-implementation.md

File metadata and controls

46 lines (34 loc) · 2.44 KB
Raw

3. Exclude RSocket implementation

  • Status: Accepted
  • Date: 2025-10-01

Context

The original project roadmap slated Phase 5.0 for the implementation and benchmarking of the RSocket protocol. Our investigation proceeded through several stages:

  1. Initial approach: Attempted to use the io.quarkiverse.rsocket Quarkus extension, but it was unmaintained and abandoned.
  2. Manual implementation: Pivoted to a manual implementation using the core rsocket-java libraries. This approach was successful on the standard JVM and passed all integration tests.
  3. Native compilation failure: The manual implementation resulted in a series of intractable, low-level native compilation errors with GraalVM. Despite multiple attempts to resolve these with build-time configuration), the issue persisted, indicating a deep incompatibility between the library's reflective features and the constraints of native compilation.
  4. CLI client failure: The standalone Java CLI client (mirroring the gRPC phase) was blocked by a low-level Netty DNS resolver conflict, which, while solvable, represented another significant deviation from a smooth implementation.

Decision

We have made the decision to exclude the RSocket implementation from this project and move directly to the final analysis phase.

The cumulative weight of the technical roadblocks—lack of a supported Quarkus extension, critical native build failures, and significant client-side tooling friction—indicates that a robust demonstration of RSocket on this specific technical stack is not currently feasible without compromising the project's core principles.

Consequences

Positive

  • Maintains project quality: We avoid introducing a brittle, difficult-to-maintain, and JVM-only implementation into a repository intended to showcase good practices.
  • Preserves project principles: Upholds our "No Shortcuts" principle by refusing to accept a solution that does not meet our standards for native compilation and reproducibility.
  • Focuses on value: Allows us to redirect effort towards the final, high-value phase of the project: a comprehensive performance analysis of the successfully implemented protocols.

Negative

  • Reduced scope: The project will not include a direct performance comparison for RSocket, which was one of the initially planned communication patterns. The final analysis will be based on REST, SSE, WebSockets, and gRPC.