Provide Pods/ShardingManager/ShardingClient implementation that relies on effect/rpc
mattiamanzati opened this issue · comments
What is the problem this feature would solve?
Given the current structure, providing implementation for non-node environments require to manually implement the lower level interfaces.
That's a shame because all those layers do is just pod-to-pod or pod-to-manager communication, so most of the orchestration of the code is around that.
What is the feature you are proposing to solve the problem?
We could provide a more general cluster-rpc package that uses effect/rpc under the hood.
That would allow to support any environment that has both platform and RPC support.
What alternatives have you considered?
We could just rely on platform/HttpServer-Client, but that means that the transport could be just HTTP.