Skip to content

refactor(runner): group route execution state - #867

Closed
afourniernv wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
afourniernv:codex/switch-1620-execution-lane
Closed

afourniernv wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
afourniernv:codex/switch-1620-execution-lane

Conversation

@afourniernv

Copy link
Copy Markdown
Contributor

What

  • Groups the algorithm, clients, runtime models, auxiliary targets, and decision targets in a private ExecutionLane.
  • Keeps caller authentication and advertised capabilities on Route.

Why

This is the first internal step toward letting a route choose between complete standard and restricted execution lanes. Keeping the execution state together avoids threading parallel sets of clients and models through the runner later.

Linear: SWITCH-1620

Notes for reviewers

This is a behavior-preserving refactor. It does not change public APIs, TOML, ownership, or per-request work.

Validation:

  • cargo fmt --all --check
  • cargo clippy -p switchyard-runner --all-targets --all-features -- -D warnings
  • cargo test -p switchyard-runner --all-features
  • cargo test -p switchyard-server --all-features

Signed-off-by: Alex Fournier <afournier@nvidia.com>
@afourniernv

Copy link
Copy Markdown
Contributor Author

Opened this too early. Closing while the design and implementation sequence are still being worked through.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant