This is a feature request for a distributed cluster router. The idea is that you want N actors spread across a cluster to be able to route to N other actors across a cluster. You want the same convenience that local routers have, namely the convenience of centralized routing logic without being blocked by a single mailbox. The state of the cluster router as well as its routing code would somehow need to be distributed and synced across the cluster.

Today this can be implemented using local routers in every node, plus custom logic to respond to cluster events and distribute routee changes to the node-local routers.

I guess the goal when we have distributed actor trees is that the cluster router should work and be synced across the entire actor tree when its distributed across many nodes.