ModelIntegrated Development of Embedded Software

1
Model-Integrated Development of Embedded Software

• CSCI 589 Eunyoung Song October 16th, 2007

Written by Karsai, Sztipanovits, Ledeczi, and
Bapty
2
Overview

• Concepts of Model-Integrated Computing
• Modeling and Metamodeling
• Model-based generator implementation
• Related Work
• Relevance to Embedded Software
• Strengths of the Paper
• Weaknesses of the Paper

3
Model-Integrated Computing (MIC)

• Based on models and generation
• Provides a flexible framework
• Fully adopts model-based development paradigms
• Introduces modeling languages
• Should be modeled using domain-specific modeling
  languages (DSML)
• Has built-in extension mechanism

4
Overview of MIC

• High-performance digital signal processing (HDSP)
• Require high throughput and high degree of
  flexibility in the architecture
• Model
• Formal structure representing selected aspects of
  the engineering artifact and its environment
• 1) The hardware architecture aspect
• 2) The signal-flow aspect
• 3) The environment aspect

5
Fig. 1 Example HDSP models
6
Model-integrated development system

• Modeling of the system and its environment, from
  multiple, interacting aspects
• Automated synthesis of design models to
  accelerate the modeling process
• Analysis of the system using analysis /simulation
  tools, which may necessitate model
  transformations
• Generation of the executable system using
  generator tools

7
Modeling (1/3)

• A. Domain-Specific Modeling Languages
• Declarative
• Use domain-specific symbols
• Have a restricted yet precise semantics
• Approaches to modeling
• Utilize a single modeling language, designed for
  software modeling (UML)
• Tailor the single language to the needs of a
  specific domain, introduce domain-specific
  modeling language (MIC)

8
Modeling Abstract Syntax
to define the data structures that can represent
our models
9
Modeling Concrete Syntax and Syntactic Mapping
Concrete syntax captures how they can be rendered
for human interaction or for machine-to-machine
communication
10
Modeling Semantic Domain and Semantic Mapping
The role of semantics is to describe the
properties of models
11
Modeling (2/3)

• B. Composition of Metamodels and Models
• Metamodel Composition
• Model Composition
• Abstraction
• Modularization
• Interfaces and ported components
• Multiple aspects
• references

12
Modeling (3/3)

• C. Tools for Domain-Specific Modeling
• Metaprogrammable generic modeling environment
  (GME)
• configurable environment to create
  domain-specific environments
• configurable visual model editor and database
  manager
• Metaprogrammable translator framework

13
Fig. 10 GME architecture
14
Metamodel

• Formal model of DSML.
• Describes the syntax, semantics and presentation
  information of a modeling language using UML, OCL
  and aspects.

15
Model Synthesis and Generative Modeling

• Compositional Modeling
• Building complex model by composing components is
  efficient modeling.
• In embedded system, many physical properties are
  not composable
• Model Synthesis
• Model synthesis can be formulated as a search
  problem.
• Fully automated synthesis is a hard problem
• Top-down engineering designing processes

16
Model-Based Generators (1/2)

• The role of Generator
• To translate models into the input language of
  analysis tools
• To translate the analysis results back into the
  modeling language
• To translate models executable into code, static
  data-structures, component configurations,
  customized generic components.

17
Model-Based Generators (2/2)

• Techniques for Building Generators
• Direct implementation the output product can be
  directly produced from the input tree
• Pattern-based approach implemented in a more
  structured way by using the Visitor design
  pattern
• Metagenerators mathematically precise modeling
  of the generators working and the generation of
  the code of the generator from that model

18
Related Work (1/2)

• Aspect-Oriented Programming
• Decomposition concept
• Provide better run-time performance
• GenVoca
• Generator technology
• Performs automated composition
• Resides between AOP and MIC
• Based on a design-space definition

19
Related Work (2/2)

• Intentional Programming (IP)
• From the conventional to a domain-specific
  activity
• Defining and implementing transformer is a
  crucial issue
• Model-Driven Architecture and Model
  Transformations
• MDA is developed my OMG.
• Specification model, implementation model
• Lends itself to transformational approaches
• Tends to utilize UML for modeling

20
Conclusion

• In MIC, DSMLs play a key role in the creation,
  manipulation, and transformation of the models.
• Metamodeling approach is extremely powerful
• Careful upfront analysis of the domain

21
Relevance to Embedded Software

• Modeling and model-based approaches play a
  central role in embedded software
• Embedded software development is based on
  domain-specific, multiple-view models used in all
  phases of the development process
• Examples of real systems illustrate how the
  model-integrated approach addresses the dynamic
  structure of embedded software.
• MIC provides a flexible framework to address
  essential needs of embedded software development.

22
Strengths of the approach

• This paper present the specific concepts about
  the model-integrated computing.
• It shows the examples to explain the modeling
  method.
• This paper shows the capability of the tools
  supporting the process.

23
Weaknesses of the approach

• Developing metamodel is not always hard
• Well-understood domain developing is not hard
• Have a possibility of reuse of the development
• Impact of Metamodel change is not always
  powerful
• Sometimes cause severe problem
• Rare possibility of metamodel change

24
References

• T. Bapty, J. Gray, S. Neema. Constraint-Based
  Embedded Program Composition. Institute for
  Software Integrated Systems, Vanderbilt
  University, Feb. 2 2001
• G. Karsai, J. Sztipanovits, A. Ledeczi, T. Bapty.
  Model-integrated development of embedded
  software. In Proceedings of the IEEE, Volume 91,
  Issue 1, Jan. 2003 Page(s)145 164.
• Model-integrated Computing. http//www.isis.vander
  bilt.edu/research/research.htmlMIC

25
Questions?