Towards Dynamically Composed Real-time Embedded Systems

Embedded systems are present virtually everywhere, from simple home appliances to complex aerospace systems, and the establishment of the Internet of Things (IoT) increases this ubiquity. Naturally, after deployment, updates might be necessary due to bugs found in the original implementation, improvement of tasks, security holes, etc. However, some domains require the deployment of a big number of devices, which often cover a vast area or are unreachable after deployment, and lead to very expensive, or even impossible updates with physical access to the devices. Therefore, remote updates are of utmost importance. In this paper, we present the existing approaches for remote updates, and analyze the overheads each one causes in the devices and in the update server. Furthermore, we present our concept for coping with dynamically composed real-time systems. Our main goal will be schedulability analysis during dynamic updates.