I am doing Research and Development to move workload from one edge device to another device. But I am stuck that how we can use k3s/k8s to actually distribute the workload across local compute ,cloud and another device . You can imagine device as a PC/ laptop for now. I am trying to develop a basic POC .

Workload – It can be like to print numbers from 1-20 using for loop ,we can print numbers from 1-10 at device1 and offload rest of the work(print 11-20) from device1 to another device2 or cloud(AWS) . For that I will have to create docker image of the workload and create kubernates cluster on cloud using Eks or run local kubernates on device2 and run pods and do compute thereby running image in pods .

Problem

i>Is my thinking to correct to upload workload in form of image to docker hub so it can be used to do compute anywhere ?

ii>How will I bring the output of the compute(Print numbers from 11-20) back to device1 and in which form ?? File / JSON / binary .

What I have tried –

#Method 1>

i>Use kubernates locally using minikube and run workload image locally on device2 . (device ->device )

ii>Also use AWS EKS to create cluster on AWS to run workload image on AWS EKS cluster in pods on cloud .(device ->cloud)

#Method 2>
Use AWS IOT green grass to shift workload .But problem here is how I can shift workload from device 1 and device 2 as AWS IOT green grass can be used between device –> cloud.

Please Help me with right direction and if there is some other alternative/technology /framework please feel free to suggest .