The Internet of Things (IoT) forms the basis for future intelligent networked innovations in the private and industrial sectors, which are also part of the German Industry 4.0 strategy. The Internet of Things consists of a number of networked IoT nodes with sensors and actuators, which are based on a large number of electronic components, some of which have extremely low energy consumption. The actual core of these devices is the software, which also guarantees the configurability and adaptability of an IoT node over a long period of time. In addition to the actual intelligent function, the software must also provide the operation of the node in the network, such as real-time operating functions, remote update and security mechanisms. The extremely limited hardware available and the high requirements in operation put a great deal of pressure on IoT software development. For this reason, the approaches used today are not competitive in terms of standard software design and embedded systems.

The COMPACT project is an industry-wide effort within the European network to develop new techniques for a fast, efficient and structured software design of extremely small IoT nodes. The focus of the project is on new mechanisms for automatic software generation for IoT nodes with extremely small memory requirements and extremely high energy efficiency. Companies along the entire IoT value chain are involved in the development process, from semiconductor companies that provide the IoT hardware platform with low-level software components to middleware providers for operating systems, tool developers, manufacturers of IoT nodes and finally developers and operators of IoT applications. The project will be carried out in the European ITEA3 network of Germany with Finland, Spain and Austria. In the German subproject, COMPACT forms a cooperation between Robert Bosch GmbH, Eberhard Karls Universität Tübingen, FZI Forschungszentrum Informatik, Kasper & Oswald GmbH, OFFIS e. V., Technische Universität München and Universität Paderborn under the direction of Infineon Technologies AG.

OFFIS is working on techniques for generating runtime and memory efficient code for simple IoT nodes. Research is being carried out to find out how these optimizations can be realized from model-based code generation and translation into machine code to the generation of binary code. Another focal point of OFFIS' research work is the support in the debugging process from the input model level down to the object and binary code. All results will be incorporated into an LLVM based compiler and debugging framework