My cloudy time machine: a scalable microservice-based platform for data processing in cloud-edge systems

Abstract

As Cloud computing is a very well developed domain, many companies tend to move their entire activity in Cloud. At the same time, there is a tendency to move some of the data processing from Cloud to Edge as close as possible to the end devices. One main advantage of this approach is minimizing the latency in communication between the end devices and Cloud. Better usage of on premise devices is also a good achievement of the Edge offload. In this paper, we propose an architecture for applications that are connected to ROBIN-Cloud or to a general Cloud. We present how we have encapsulated Python-based microservices in Docker containers. We provide an implementation for the My Cloudy Time Machine application - GIGEL (Guided Intelligent GEared Legend), a nearby autonomous assistive robot. We use this prototype to evaluate the scalability of the proposed architecture. We also present results that show how to gain high performance by tuning a container-based embedded system.