Last modified on June 16, 2008, 4:06 pm

Workshop instructions
 
Program
Participants
Important dates
Local information
Organization

Supported by:

PROGRESS Workshop on
Component Models for Embedded Systems (COMES'08)

June 17th - 18th, 2008

Component-based development aims of separation of development of components from systems, and reuse of existing components. This approach brings many (known) advantages, but also many challenges of which some are specific or in particular important for embedded systems. Since for embedded systems non-functional properties and resource constraints are of particular importance, this workshop focus on component-based approaches that ensure predictability of the system properties. Predictability is a quantified measure of confidence and precision that a particular property (or property value) will be achieved. The aim of the workshop is to
  • Present and discuss the current research and practical results in development of embedded systems using component-based development approaches.
  • Discuss and point out the challenges and possible solution directions in applying the component-based approach to achieve predictability of component-based embedded software systems.
The workshop is setup as a set of session in which each session will focus on particular challenges. The sessions will start with some introductory presentations and will continue with discussions, and hopefully lead to some conclusions.

The list of challenges of interest:

1 - System properties vs. component properties

In embedded systems many properties (time properties, dependability, etc.) are to a large extend determined by the system requirements and constraints. The question is how these properties are related to component properties.

  • For a given set of system quality attributes, which attributes are required of the components concerned?
  • To which extent, and under which constraints, are the emerging system attributes determined by the component attributes?
  • How can the system quality attributes be accurately predicted, given that the component quality attributes are known with a certain accuracy?
2 - Component models for embedded systems

What are the characteristics of component models for embedded systems:

  • What concerns (and solutions) should be built in component models?
  • Which architectural solutions are suitable for different classed of embedded systems?
  • How can component attributes that are dependent on the underlying structure be specified/modelled?
3 - Generality and efficiency

To be reusable, components have to be general enough. Generality can however decrease the system performance. Important questions include:

  • How to build component models (and components) that provide efficient code (with respect to timing characteristics, resource usage), enable reusability, and allow/facilitate prediction?
  • How to design component models that enable software development separated from hardware development?
  • To which extent should components be aware of the system context?
  • How to combine diversity (for example reusing components on different underlying platforms) and efficiency
4 - Development and run-time environment

In embedded systems the development environment is usually different from the run-time environment; typically the target systems are quite different from the development platforms. This requires support on both platforms.

  • What are the deployment policies (run-time update or, building up the entire system, enabling on-line optimization of resource utilization, etc.)?
  • What are the methods that efficiently comprise specifications of different abstraction levels and different concerns (for example, how to unify deign of a system and design of a component, or system properties with component properties)?
  • How to efficiently integrate a chain of tools needed for a sufficient development, monitoring, verification and maintenance of component-based embedded systems.
  • How to ensure the system integrity by dynamic deployment of components?
5 - Particular attributes and their composition

Theories for how to model and compose attributes of particular interest for embedded systems: timing attributes (response time, execution time,), performance (resource utilization, throughput), dependability (reliability, safety, integrity, etc.), as well as how to provide a consistent framework for reasoning about multiple of these properties.

Workshop format

Each participant (or possible group of participants in case of multiple attendants from the same group/organization) is required to prepare a brief position statement no later than May 9. This statement should

  • state the focus and interest in relation to the above challenges
  • indicate 1-2 challenges of primary interest
  • for each of the indicated challenges
    • outline a 10-30 min presentation specifically addressing the challenge or a closely related aspect (an abstract is sufficient)
    • provide 1-3 articles of relevance
For each challenge a chair and a rapporteur will be appointed. It will be the task of the rapporteur to summarize the conclusions of the session, briefly at the end of the session, and also in some more detail as an introduction to the panel discussion that will conclude the WS. It is also the ambition to compile (and possibly publish) a report summarizing the findings of the WS.
Web page hosted at Mälardalen Real-Time Research Centre (MRTC),
Mälardalen University, Sweden. Webmaster: Thomas Nolte
This page has been accessed 336 times.