The design and engineering of solutions based on sensor data is currently based on legacy low-level sensor communication. The transition to IoT-based sensors has started to provide IP-based sensor communication based on service concepts. This requires radically different solution architectures and the associated engineering process. Here, IoT sensors will provide services instead of a communication interface, as a means of integrating into application solutions. It is further anticipated that the size of the solution will reach well beyond that of the current legacy automation implementations.This paper addresses architectural concepts for very large-scale SoS- and IoT-based solutions. Architecture concepts and their associated engineering concepts are considered. IoT and SoS evolvability, run-time dynamics, scalability and segmentation, run-time engineering, self-engineering, self-mitigation, and machine-to-machine business are some examples of such concepts. These concepts must be coupled with the requirements such as the management of IoT/SoS solution functionality, security, safety, maintenance, and evolution.