Logotype Mälardalenn University
This web site has been moved.
Please visit our new and improved web site at: http://www.es.mdh.se/
Startpage www.mdh.se

The Platform/Deployment Research Direction

Researchers involved

Mikael Sjödin (Professor), Dag Nyström (Senior Lecturer), Jukka Mäki-Turja (Senior Lecturer), Andreas Hjertström (PhD Student), Rafia Inam (PhD Student), Marcus Bohlin (PhD Student), Kaj Hänninen (PhD Student).

Context and research overview

The Platform (later renamed to Deployment) Research Direction has developed novel technologies for platforms for execution of component-based software and has developed deployment techniques to allow predictable deployment of ProCom-components to physical and virtual platforms. Platform related techniques include design-time and run-time technologies for data-management, operating systems support for predictable and resource-efficient execution of components, and analysis techniques to bound resource usage of systems utilizing these techniques.

Major results

Highlight 6 - Deployment model for reuse and run-time platform [Show]

We created a model for separating the models of components and their implementation [21]; the model allows reuse of both component-models and component-implementations. Further, the model allows a 2-step deployment process; where components are first deployed and validated on a virtual architecture and secondly the virtual architecture is mapped to a physical architecture. The second mapping is done without altering the timing behavior of the virtual architecture; thus minimizing integration problems in the time-domain.

To support the above 2-step deployment model we have implemented code-synthesis and a run-time environment to support execution of the virtual architecture upon a physical architecture [22]. The run-time environment builds on our extension to FreeRTOS and executes on an AVR32-platform.

Our approach to execute virtual nodes (or virtual architectures) is based on a 2-lever hierarchical scheduling framework (HSF). The HSF has been implemented as a stand-alone extension for FreeRTOS with support for hard real-time analysis and predictable synchronization between virtual nodes [23, 24]. This implementation is distributed as an open source extension to FreeRTOS and it also support execution of legacy FreeRTOS applications within the HSF.

Highlight 7 - Data management in embedded systems [Show]

We have devised a modeling technique to describe global data in distributed embedded systems [25, 26]. The technique is based on observed deficiencies in state-of-practice with respect to data-management in the design-process of industrial embedded systems. The modeling technique is implementation agnostic; specifically implementations could be made with or without a real-time database management system.

The concepts of Component Based Software Engineering (CBSE) and RTDBMS (Real-Time Database Management Systems) are contradictory. CBSE promotes encapsulation and prevents dependencies to external objects (such as global data). RTDBMS provides flexible, predictable, safe and secure sharing of global data. Both concepts have potential to lessen the burden of implementation of embedded systems. To resolve this contradiction we have developed a “database proxy” concept, which support both design-time and run-time management of data-entities and their coupling to component-ports [27, 28].

  • Mälardalen University |
  • Box 883 |
  • 721 23 Västerås/Eskilstuna |
  • 021-101300, 016-153600 |
  • webmaster |
  • Latest update: 2013.02.25