Edge processing and exchangeable secure transport layer for IoT devices

Offentlig prosjektsammendrag

qbee.io is a device management and remote access platform for embedded Linux edge devices. In discussions with customers we get frequently asked if we could handle the secure transport layer from the application platform to a cloud backend as well. In this project we want to investigate if it is feasible to offer a flexible and exchangeable multi-cloud architecture for IoT application data such that cloud backend platforms can be exchanged. This way all commercial cloud providers are supported as well as open source solutions or the up-coming pan-European GAIA-X platform. By providing a simple UI we allow to hot-swap cloud services on edge devices thus breaking vendor lock-in and regulation challenges. In order to achieve this data needs to flow through a standard queue on the edge device. This queue will have modules to map data and connect to different cloud connectors. The qbee platform will connect to cloud key management systems and provide a basic key management for open source transport layer solutions. Since the application data will flow in the qbee messaging queue anyhow this project will also research the feasibility to do general time-series anomaly detection on the application data through machine learning. A lightweight machine learning algorithm will be run as a plug-in module in the message queue while coefficients are calculated and optimized on a digital twin in the backend. New coefficients will then be delivered to all edge ML modules on a continuous basis through the existing qbee agent. If successful this will allow to retrofit IoT application with basic anomaly detection.