|

Capella and ARCADIA at Thales: Should tool and method be coupled?

Capella and ARCADIA

Much has already been written on these pages about Eclipse and already reported on the working group PolarSyswhich deals with the development of embedded systems with Eclipse-based tools. Today we are talking about Capella, a tool from the PolarSys toolbox, and a method called ARCADIA. Two things are of interest to the reader here: Firstly, Capella and ARCADIA are used by Thales and, according to Thales, is already being used productively and successfully throughout the Group. On the other hand, it is about the close coupling of tool and method, which is also rather unusual. What this approach means will be explained below.

Next week Tuesday, on June 27 at 19:00 (CEST)together with the companies Jama Software and Intercax, where we will present an integration between Jama (requirements management) and Cameo (SysML modeling) show. Register for free >>

By the way, you can read more about what has been said here in the recently published article Coupling Methodology and Tooling for Systems Modeling: ARCADIA and Capella at Thales can be read here.

Open Source

Capella was originally developed as an internal, proprietary tool at Thales. It was only made open source in 2014. In the years since its release, the community and the ecosystem have grown considerably, which shows that the calculation has worked out: both for Thales and for the many users.

There are now many factors that speak in favor of open source in the commercial environment. The financial costs are almost irrelevant. It is much more about sustainability, long-term availability and independence from software manufacturers. Capella, as well as accompanying documentation and sample data, can be downloaded free of charge from the PolarSys website can be downloaded. PolarSys is an Eclipse Working Group, and accordingly Capella is under the auspices of the Eclipse Foundation.

ARCADIA

ARCADIA stands for "ARChitecture And Design Integrated Approach" and covers method, abstract ontology and views. A special notation is introduced for this purpose. The notation is one of the reasons why a separate tool was developed.

ARCADIA itself is an approach to modeling hardware and software with the aim of understanding customer benefits and communicating the system architecture. It is also intended for the coordination of interdisciplinary work. Verification and validation (V&V) should be addressed at an early stage. To achieve these goals, the approach follows several principles, such as functional analysis and separation of need and solution.

Concepts and model types

ARCADIA has several semantic concepts that can be displayed in various diagram types in Capella. These include

  • System Dataflow Diagram for the functional analysis
  • Class Diagram for data analysis
  • System Architecture Diagram for the analysis of behavior
  • Physical Architecture Diagram for physical partitioning

As with many modeling approaches, the diagram in Capella is merely a view of the model. Accordingly, the model allows the elements to be linked together, with ARCADIA determining the procedure. The model also has other constructs for dealing with the model, such as "replicable elements", to enable reuse.

In addition to the diagrams, there are also perspectives. These are views that are particularly suitable for certain activities. These are

  • Operational analysisto capture the needs of the stakeholders
  • System Need Analysisto capture the system's contribution to solving the problem
  • Logical Architecturewhich describes how the system will solve the problem
  • Physical Architectureto capture the physical manifestation, i.e. the tangible system
  • Product Building Strategyi.e. how the system should be configured and integrated.

Capella

After existing modeling languages such as UML and SysML have now become widespread, it is initially surprising that a new language - and tool - had to be developed. And this approach was also tried out in the early days of ARCADIA: In the 2000s, attempts were made to use profiles to adapt SysML for use with ARCADIA. However, it turned out that this approach was very complex. The tools were not easy to configure and users complained about usability.

Capella was inspired by SysML, but is a language in its own right, which makes it much simpler. Due to its specialization, Capella is not as flexible as SysML, but it is easier to understand and more powerful. The in-house development also enabled closer integration with the method, which makes it easier to use in many places and guides the user. For example, Capella uses specific colors for different perspectives.

When working with the model, the interlocking between tool and method also comes to the fore by guiding the user and automatically supplementing the model. A good example is working with interfaces between components, which can be initialized accordingly when linking.

Capella is a semi-formal language, which in many places in the tool supports the user in nerve-racking activities, such as maintaining dependencies, model generation, automated "quick fixes", etc. Many of these functions may remind readers with programming experience of the powerful features of modern programming environments.

Scaling

The concepts just described are necessary for scaling, and scaling is essential for industrial use. Of course, there are many aspects to scaling: it is not just about the size of the model, but also about spreading the method and tool in a large organization. A simple language (at least in comparison with SysML, for example) is also helpful for this. Restrictions have been introduced in the modeling language that ultimately benefit scaling. For example, only functions that do not contain any sub-functions can have inputs and outputs. On the one hand, this enables a quick completeness analysis and, on the other hand, nested inputs and outputs can be automatically carried out over several levels. The article cited above contains some examples of this.

Conclusion

The decision to use any combination of methods and tools should be strategic, and this also applies to Capella and ARCADIA. In this respect, careful consideration should be given to whether there is a business case and whether a return on investment can be realized. However, Capella is attractive for several reasons: the many years of support and use by Thales give the approach a certain degree of stability. The outlay for a pilot is limited. And there is now a strong community that continues to develop Capella and helps with questions.

Personally, I find the close link between method and tool very interesting. I would love to carry out a project with Capella and ARCADIA.

Do you already have experience with Capella and ARCADIA? Then share it in the comments section!

Image: PolarSys/Capella

Similar Posts

4 Comments

  1. I have mixed feelings about the dovetailing of methodology and tools. On the one hand, this provides users with guidance, on the other hand, this interlocking imposes restrictions. This interlocking has been common practice in the enterprise architecture environment with TOGAF/Archimate tools since the beginning and is fully accepted in this environment.

    In the technical environment, however, there is no generally recognized methodology like TOGAF in the enterprise architecture environment. In addition, the methodology is discussed and questioned much more in the technical environment. Accordingly, acceptance is likely to be more difficult in the technical environment. In my experience, UML tools that specify a methodology (e.g. MID Innovator) were generally not seen as helpful.

  2. I have been observing ARCADIA/Capella since December 2014 and have been working intensively with both the methodology and the tool.

    Methodology:
    The methodology is the best I have seen so far for large projects. The methodology is much more lightweight than NAF v3, for example, but still offers all the necessary levels of abstraction. Furthermore, the sequence of views makes sense.

    To the tool:
    The tool works well. Due to the smaller number of ARCADIA elements and their attributes compared to SysML, the menus are also clearer than with SysML tools.

    Overall, I find ARCADIA more pleasant than SysML in a pure systems engineering environment. However, I see the transition to software engineering as a problem. UML is the de facto standard in software engineering. However, a transition from ARCADIA to UML is a media break. With tools that interpret SysML as a UML profile, the transition is possible without media discontinuity.

  3. From our experience in using Capella/ARCADIA on control and monitoring systems for railroads, we can only emphasize this. The clear separation into the different levels of abstraction helps both beginners and experts to structure their systems better. You avoid the otherwise often occurring mixing of logical and physical views and do not lose the overview so easily in larger systems.

  4. Info about ARCADIA/Capella: On Tuesday, March 13, there will be a free Capella Day in Stuttgart.
    Capella Day is an opportunity to meet members of the Capella ecosystem and learn from their expertise.
    Deepen your knowledge of the Capella Workbench, discover the latest plugins and extensions, such as document generation, collaborative working or simulation.
    For further information or to register on the list of participants https://wiki.eclipse.org/Capella_Day_Stuttgart_2018

Leave a Reply