Hybrid Serverless Processing in Environmental Monitoring - Building Digital Twin

This paper discusses the application of a cloudedge continuum implemented with containerized and event-driven processing for environmental monitoring. We propose an architecture based on a serverless computing paradigm that allows managing software and digital assets. We demonstrate its functionality on a levee monitoring use case. In the use case, we utilize CloudEvents specification to deliver messages in a protocol-agnostic way. We also build and evaluate mechanisms for task migrations based purely on Kuberentes and Knative. Because the execution process spans across different serverless environments, we utilize hybrid servereless approach 1. In our system prototype, we also build and evaluate the concept of nanoservices 2 in delivering environmental monitoring solutions. Our early results show that, with our approach, we can build scalable and elastic digital twin applications and scenarios capable of analyzing large-scale infrastructure.