Usage of ray on edge devices

Hi everyone,

I’m currently developping an edge device solution and I came accross ray as a framework to execute processors and algorithms on collected data.
I was therefore asking myself whether Ray could be used for such an application, considering the idea would be not to deploy a complete Ray cluster but rather some kind of “local” instance. On my device is a k3s cluster which I also use for other components. My idea would be to have ray inside a pod (with ray.init() without any address launching a local ray instance inside the pod). Has anyone any thoughts about that ?

Thanks in advance

P.S. I saw the

Is Ray suitable for edge real time processing?

topic and asked the creator of this topic but never got any response that’s why I’m trying with a new topic :slight_smile: