Architectural Modeling of Embedded Systems (AIMES)


Key words/Classification:

Software Architecture, Architectural Design and Modeling, Architectural Evaluation, Architectural Process, Distributed Systems, Embedded Systems, Performance

Summary

Architectures become increasingly important as a tool for managing the complexity of IT products and their development process. This particularly applies to Embedded Systems with their complex hardware/software interaction, reactive behaviour and their high reliability demands. This research aims at improving the quality of software-intensive Embedded Systems and their development process. After a thorough analysis of the requirements, the following research questions were formulated in co-operation with the industrial partners:

  1. How can the performance (response time, throughput, resource utilisation, etc. of architectures be evaluated? Since Embedded Systems are resource con-strained, answering this question is extremely important. In order to check the feasibility of a particular architecture, it must be decided in an early stage whether the performance requirements can be met or not.
  2. How should architectures be modelled? In order to assess the performance of an architecture, it must first be adequately described. Within the framework of this project a component-oriented approach was chosen since component technology is important for promoting software reuse and for reducing devel-opment costs.
  3. How should architectures be developed, i.e. how should the architecture proc-ess look like? In order to tackle this question, also an interesting paradigm was chosen, namely the use of scenario techniques for improving the quality of architectures. Scenario techniques can be used during the early architecting stages (requirements engineering) as well as during the evaluation of the ar-chitecture. During the preliminary stages, the main goal is to promote the co-operation with and between the interested parties (stakeholders). During the evaluation, the goal is to limit the number of changes induced by the evalua-tion of the architecture) by anticipating them already in the requirements engi-neering phase.

The answers to these three questions will be incorporated in a design method for the architecting of Embedded Systems and a number of supporting tools. This method and the accompanying tools are also an important means for transferring the results of this project to the industry in general, and to the industrial partner in particular.

In order to achieve these results, a number of architecting case studies will be per-formed together with the industrial partner. This research is therefore problem driven and not theoretical. To promote the exchange and consolidation of ideas and to ex-ploit the synergy between industry and academia, the three junior researchers, their industry coaches and their TU/e coaches work closely together. The multi-disciplinary nature of the architecting process is reflected in the individual industrial projects and in the fact that two TU/e faculties are involved.

People

Ph.D. students

Staff

Funding

Industrial partners

In cooperation with:

Research papers

Project information