Skip to Main Content
Cost pressure, flexibility, extensibility and the need for coping with increased functional complexity are changing the fundamental paradigms for the definition of automotive and aeronautics architectures. Traditional designs are based on the concept of a Federated Architecture in which integrated hardware/software components [Electronic Control Units (ECUs)] realize mostly independent or loosely interconnected functions. These components are connected by bus and cooperate by exchanging messages. This paradigm is now being replaced by the Integrated Architecture, - the concept comes from Integrated Modular Avionics (IMA) introduced by the avionics community (see C. B. Watkins and R. Walter, ?Transitioning from federated avionics architectures to integrated modular avionics?, in Proc. 26th Digital Avionics Syst. Conf., Oct. 2007) but it is certainly general and applicable to other fields and in particular, automotive - in which software components can be supplied from multiple sources, integrated on the same hardware platform or physically distributed and possibly moved from one CPU to another without loss of functional and time correctness and providing a guaranteed level of reliability. This shift will decouple software design from the hardware platform design and provide opportunities for the optimization of the architecture configuration, increased extensibility, flexibility and modularity. However, the integration of software components in a distributed system realizing a complex functional behavior and characterized by safety, time and reliability constraints requires a much tighter control on the component model and its semantics, new methods and tools for analyzing the results of the composition, whether by simulation or formal methods, and methods for exploring the architecture solution space and optimizing the configuration. We provide a general overview of existing challenges and possible solutions to the design and analysis problem, with - - special focus on the automotive domain. The development of such methods and tools must necessarily consider compatibility with existing modeling languages and standards, including UML, AUTOSAR and synchronous reactive models, on which the widely used commercial products Simulink and SCADE are based.