Grappling with complexity is a challenging task. But that is the world of systems engineering. The problems we face are complex, and they demand complex solutions. Managing that complexity in a way that solves the complex problem while keeping all of the aspects of the problem and potential solutions visible and accessible to the design team is the crux of the challenge. The team needs to grasp the effects of changes to the solution as they are made. They need to be able to communicate the current state of the solution to stakeholders and team members alike. This can only be done using an integrated and complete model of the system design.
The model that underlies the system design in model-based systems engineering (MBSE) is a very powerful tool. That model is created by populating a metamodel with relevant information about the design of interest. The metamodel defines the set of classes, relationships, and properties (also known as attributes) that will be used to construct the model. Put more simply, the metamodel provides the language used to describe both problem and system, a schema is the instantiation of that metamodel into a database so that it can receive the design information according to the established rules and definitions, and the model itself contains the design information organized by the schema. The metamodel allows the model to be understood and communicated without ambiguity. This makes the model a powerful tool.
With the model, the designers can track the progress of the design, maintain consistency, communicate the design specifics to others, and trace the impact of changes and alternatives. But, if there is no metamodel to define the model structure, these benefits become exponentially harder to realize.
Use of a single database
Using a single, integrated model allows the design team to surface their ideas and choices and make these visible to all. Attempting to do this by tracing from model to model across different databases is inefficient and ineffective. The responsibility for consistency and completeness is always levied on the designers. Breaking a model across multiple databases levies the bookkeeping of consistency on the designers as well (as opposed to an integrated model with guaranteed consistency), and the power of the integrated model to ensure the design integrity is lost to them.
Practices that increase risk
Despite the risk of losing the power of the model, practices that have exactly that result persist in the systems engineering community. The most common failure in this regard is the use of tools that do not rest on a complete, integrated metamodel or even have any underlying system model at all. These tools fail to relate the entities used in the design to each other across the borders between the tools. For example, a “requirements” tool may handle all the aspects of requirements capture and management but have no way of relating them to functions and no way to address the functionality of the system design at all. It is not enough to cobble such incomplete models together in an effort to achieve a “system” model, because the power to track and assure consistency across domains will never exist without the connections of a single model.
Another brand of risk is created by attempting to model systems using tools that compartmentalize the model at the views level. In this case, any given view will contain only that subset of information about the design that the view definition calls for. This means that the design is described in a set of snapshots, each offering some detail but without anything approaching the level of a comprehensive picture. Without building the model based on a metamodel that defines all aspects of the system, there is no truly complete model to track the design. It is crucial that this model incorporate all of the elements of the system design including the relationships. Data dictionaries – either capturing the views themselves or capturing the entities on the views – aren’t sufficient to expose and manage the relationships that are the key to understanding how a system will behave. Ultimately it is the system behavior that will satisfy the requirements that drives its creation. Understanding that behavior is the key to designing effective solutions.
Benefits of a robust tool
A good system design tool uses the rules for producing a particular view to portray a subset of the design information. The view is produced from the model and offers an accurate depiction of that portion of the model. Where a design tool is also capable of capturing changes in the views as changes to the underlying model, the views become a way of communicating design information into and out of the model. A truly robust design tool facilitates this two-way communication.
Direct simulation and verification
When the model is based on an executable metamodel, it is possible to simulate the system behavior directly from the model. Because the simulation is based on the well-defined semantics, the system behavior can be simulated with a high degree of integrity. In doing so, the system model becomes a virtual prototype of the system from day one of the project, increasing in fidelity as the model matures. Such a simulation minimizes the risk that would be created by verifying the system design based on an interpretation of the system specification. The specification and the simulation are generated directly from the model itself.
A powerful tool
A well-designed model based on a strong metamodel is a powerful tool for understanding, creating, and managing a system design. It contains all the relevant design information as conceived by the designers. Its completeness and consistency can be assured throughout the design process. Views drawn from it provide a real-time look at the state of the design at any given point. A robust tool producing and maintaining such a model is the sine qua non of the power of the model.
If you would like to see such a tool and experience the real power of a true system design model, don’t hesitate to contact us here. GENESYS and CORE stand ready to leverage the power of your designs.
