Software has become the driving force in the evolution of many systems, such as embedded systems (especially automotive applications), telecommunication systems, and large scale heterogeneous information systems. These so called software-intensive systems, are characterized by the fact that software influences the design, construction, deployment, and evolution of the whole system. Furthermore, the development of these systems often involves a multitude of disciplines. Besides the traditional engineering disciplines (e.g., control engineering, electrical engineering, and mechanical engineering) that address the hardware and its control, often the system has to be aligned with the organizational structures and workflows as addressed by business process engineering. The development artefacts of all these disciplines have to be combined and integrated in the software. Consequently, software-engineering adopts the central role for the development of these systems. The development of software-intensive systems is further complicated by the fact that future generations of software-intensive systems will become even more complex and, thus, pose a number of challenges for the software and its integration of the other disciplines. It is expected that systems become highly distributed, exhibit adaptive and anticipatory behavior, and act in highly dynamic environments interfacing with the physical world. Consequently, modeling as an essential design activity has to support not only the different disciplines but also the outlined new characteristics. Tool support for the model-driven engineering with this mix of composed models is essential to realize the full potential of software-intensive systems. In addition, modeling activities have to cover different development phases such as requirements analysis, architectural design, and detailed design. They have to support later phases such as implementation and verification and validation, as well as to systematically and efficiently develop systems.
TopIntroduction
Software has become the driving force in the evolution of many systems such as embedded systems (especially automotive applications), telecommunication systems, and large scale heterogeneous information systems. These so called software-intensive systems are characterized by the fact that software influences the design, construction, deployment, and evolution of the whole system (cf. Recommended Practice for Architectural Description of Software-intensive Systems, IEEE-Std-1471-2000 (2000)).
A good example for the growing importance of software in many advanced engineering fields is automotive software. It has become an important factor in the development of modern high-end vehicles where today the size of the software grows at an exponential rate. About 70% of the innovations in these cars today are software driven and the percentage of costs due to the development of software is expected to increase from 20-25% up to 40% in the next years (cf. Grimm, 2003; Hardung, Kölzow, & Krüger, 2004). In addition, today more and more functions can only be realized by the cooperation of different control devices. This result in dramatically increasing complexity (cf. Grimm, 2003) compared to the separate control devices employed traditionally which had to fulfill a single task only.
Furthermore, the development of these software-intensive systems often involves a multitude of disciplines (cf. Figure 1 (a)). Besides the traditional engineering disciplines (e.g., control engineering (CE), electrical engineering (EE), and mechanical engineering (ME)) that address the hardware and its control, the system often has to be aligned with the organizational structures and workflows as addressed by business process engineeringbusiness process engineering (Wirsing, 2004). The development artifacts of all these disciplines have to be combined and integrated in the software. Consequently, software engineeringsoftware engineering (SE) adopts the central role for the development of these systems.
Figure 1. Software-intensive systems w.r.t classical domains (cf. Figure 1 (a)), and the assignment of phases to disciplines (cf. Figure 1 (b))
The development of software-intensive systems is further complicated by the fact that future generations of software-intensive systems will become even more complex and, thus, pose a number of challenges for the software and its integration of the other disciplines. It is expected that systems become highly distributed, exhibit adaptive and anticipatory behavior, and act in highly dynamic environments interfacing with the physical world (Wirsing, 2004). Consequently, modeling as an essential design activity has to support not only the different disciplines but also the outlined new characteristics.
Tool support for the model-driven engineering, with this mix of composed models, is essential to realize the full potential of software-intensive systems. In addition, modeling tasks have to cover different development phases such as requirements analysis (Req.), architectural design (Arch.), and detailed design (Design). They have to support the phase specific modeling and analysis tasks as well as the synthesis and analysis tasks required for later phases such as implementation and verification and validation (V&V), as well as to systematically and efficiently develop systems (cf. Figure 1 (b)).
The integration of the different models used in the different disciplines is of major importance for the systematic development of software-intensive systems. Therefore, in this chapter we focus on the modeling techniques of the different disciplines involved as well as their proper integration (cf. Giese & Henkler, 2006).
We consider the different modeling options in the three main disciplines to reflect the different modeling techniques employed for software-intensive systems: UML (Unified Modeling Language), the de facto standards for software engineeringsoftware engineering, MATLAB/Simulink, the de facto standards for traditional engineering, and a number of approaches for business process engineeringbusiness process engineering.