Home Search AA

Importance of embeeded systems


As the embedded systems landscape – with an annual production of over nine billion computers, several important trends become visible:

  • Rapid technology development, with micro-and nanoelectronics becoming smaller, cheaper, faster, more power-efficient, and with increased use of both wired and wireless connectivity
  • Integration of products into larger, more complex and increasingly heterogeneous systems
  • The application of embedded systems technology in new products, with developers who increasingly tend to be application domain experts rather than experts in computer technology.

Two crucial EU programmes – FP7 and the Competitiveness and Innovation Programme (CIP) – both give high priority to Information and Communication Technologies as drivers for competitiveness. Embedded systems are one of the strategic research priorities of the new Information Society Technology Priority in FP7.

Inspired by other engineering disciplines, there is a trend to build software system from preexisting components that in an efficient way can by integration of software components. The aim of DICES is to utilise this approach for development of embedded systems software.





The overall objectives of the project include modelling distributed embedded systems as component-based systems that will enable better reusability of software components. At the same time the goal is to provide methods and develop tools for prediction and optimisation of certain system properties (such as performance, and resource usage).


A combination of reusability, efficiency and predictability is a novel part in the research and can lead to innovated results – new or adapted component models that enable predictability by construction.

To achieve the overall goals the following objectives will be addressed in the project:

  •  Provision of detailed state of the art of component models and service standards, including the challenges related to the predictability of the system properties, and a detailed state of the art of quality modelling, composition and analysis of quality attributes relevant for embedded systems.
  • Adoption of some service-oriented and component-based technologies to improve the predictability. The significance of this is possible improvement of system qualities and increased development efficiency due to reusability and decreased requirements on exhausted testing.
  • The adoption is tightly related to models of quality attributes and transformation between description of constructive parts, i.e. architecture and possibly detailed design, and elements in the theories of specification and composition of quality attributes.

The domains on which the results will have the most significant impacts are: Embedded systems with limited resources – for example automotive industry, consumer electronics, small embedded devices.



Validation in an industrial context


The approaches will be applied to real cases in which the metrics for the efficiency, reusability and quality will be defined and the measurements based on these metrics will be performed. This part is very important as it will demonstrate in a quantitative manner the feasibility of the approach.

In particular the industrial case study IForestFire will be used as a demonstrator for improved characteristics. The system will be partially rebuilt using new technology and the following characteristics will be evaluated:

  • Predictability:

i. Resource usage – by system construction it will be possible to calculate and optimise the use of memory, and measure the real use and compare it with the predicted values.

ii. Software reliability – a model for calculating reliability of the installed software and possibilities for increasing reliability will be demonstrated.

iii. Real-time characteristics – response time of computational units and communication units will be modelled, predicted by the theory andmeasured.

  • Reusability

 A reusability model and measurements will be provided for different control stations (percentage of software reused and percentage of software modified). This will be compared with the old solutions.



Approach and methodology


The approach follows a research path characteristic for software engineering: It starts from recognition of problems relevant in an industrial context and identifying the final goal that leads to a problem solution. The real industrial problem is transformed into a research context in which the research questions, approaches and possible solutions to the research questions are provided. The research results are validated in the research context, and then in the real industrial context.


Dices project approach


This approach includes the following methodologies:

  1. Analysis of an industrial case study to gathering the concrete and detailed requirements, constraints and concerns, and to understand the essence of the problem.
  2. Building knowledge – studying state of the art and state of the practice in order o understand the theoretical bases of methods that might be utilised in the problems solution.
  3. Creating theoretical models using basic theories in computer science, software engineering, component-based software engineering, web-based technologies to provide the theoretical and principle solutions.
  4. Building tools prototypes for modelling components, their parameterisation and system configuration to achieve predictable behaviour of the systems.
  5. Validation of the model and tool by applying on the existing real industrial case iForestFire, building a new version, comparing its properties with the old version and providing directions for further improvements.
  6. Generalisation, dissemination, and possible exploitation and commercialisation of the results in a broader industrial context –in this case application of component-based approach to increase reusability and development efficiency and enabling predictability of the system properties. Examples of the companies that can utilise the results are in Croatia Ericsson and Končar, or in Sweden ABB and CC-System (a car control-device development company)