Example required

[kinderscala]

Dear contributors,

It does look interesting, could you share some basic snippet-example for using ConHub for the case
when it should manage Con on separate node? Something like

---

| EDGE Server (Registers Remote Con) | <----->. | Client for remote connections (Lookup) |

[t3hnar]

Hello, could you elaborate on word manage in the given scope?

[kinderscala]

Thank you for the quick response,

we are building application to manage set of robots, where the core is placed in cloud and to connect to the robots we re using ZigBee protocol, we want to register every connection on edge server (because of hardware socket) and notify application server that connection has been registered - unregistered.
Application should send control commands to subset of robots and get the feedback sensor commands asynchronously back.
As you can see there are two components:
Edge Server that manages socket connections and remote computation engine, that should get access to those connection and has several nodes (we use Akka Cluster/Persistence).
We are looking into conhub as an alternative to akka grpc.

Val

[t3hnar]

Oh, that's familiar problem :)
I've just added additional sentence to README

indeed ConHub could fit here, looks like following features that you might need are already supported by ConHub

• you can register callback, which will be called upon new connection registered somewhere in the cluster
• you can send message to your robots from anywhere in the cluster, even more, you can send same message to multiple robots by adjusting lookup criteria

[t3hnar]

Client for remote connections (Lookup) |

ConHub does not cover scope of the way you implement connections. From ConHub perspective that could be just a function (Con, Msg => Unit). ConHub should be launched on all nodes and that's it. Instance of ConHub at the same moment is a registry as well as an endpoint to route messages to connections. It's up to user to implement function used by local conHub with local connections