A new approach to Mission Critical System design: Model Based Cybertronics Systems Engineering

Complexity drives change

The complexity of digital functionality within cyber-physical systems is skyrocketing, stretching the utility of traditional design methodologies. These methods no longer scale, critical architecture decisions are made on a “guess-timate” basis and, even then, are often made too late in the design process. Even the electronic units – cybertronics subsystems – are a multidisciplinary undertaking, as they include different implementation domains ranging from compute enclosure to chiplet design and multiple software stacks.

Mastering this complexity requires a new model-based design methodology that can manage all required design domains, deep subsystem hierarchies, while incorporating the varying perspectives of the design teams. It must also manage the requirements allocation as well as verification threading in the project’s design and integration phases. The Model-Based Cybertronics Systems Engineering (MBCSE) methodology outlined in this blog is designed to tackle the challenges of cybertronic design, namely:

• “Multi-disciplinarity”. Cybertronic implementations consist of multiple physical implementation options (e.g., PCB, 3DIC, SOC or FPGA) and SW running on diverse types of computing platforms. Existing methodologies struggle with the management of the multi-disciplinary nature of the development of systems.
• The allocation and propagation of requirements throughout the system design process.
• Complexity. Cybertronics system with complex SW running on even more complex HW with limited data processing and transfer capacity cannot be specified using documents. This requires an abstracting methodology allowing the increase of model fidelity while moving down the subsystem tree.
• The allocation of functions through to the appropriate operational components.
• Verification across multiple implementation domains using different verification methodologies.

MBCSE : Methodology inside a flexible framework

Our new MBCSE methodology is a combination of multiple methodologies and tools necessary to create an overarching framework.

In it, the cybertronics systems engineering process is requirements-driven with parametrized requirements for automated verification. The requirements capture and management can be done using any requirements management tool that manages parametric attributes of the requirements. The requirements and parameters are linked to the threading database. The architecture modeling is itself complex, containing system analysis, system exploration and the implementation specification of the current system level. To enable success, the modeling methodology needs to allow flexibility while enforcing consistency so that the architects of different levels of the subsystem’s hierarchy will follow the methodology. To fulfill the methodical requirements, MBCSE uses concepts from two well-known systems engineering methodologies: Arcadia and Property Modeling Methodology (PMM).

Arcadia is a graphical, layered system modeling methodology in which every layer is a design step with different focus areas: functional modeling, architecture exploration and physical implementation. A clear separation of functions from structures enables flexible allocation of functions to different structural elements (e.g. programmable devices, dedicated accelerators, or standard components). Arcadia also supports recursive system decomposition with subsystem transition. This approach allows the allocation of functions to subsystem components and decomposing them in a separate project, ultimately enabling a thorough exploration of the subsystem without either overloading or polluting the upper-level system model.

Properties are used to insert additional information to the system model in which it is needed. Properties can be requirement parameters, algorithm attributes or implementation constraints (or whatever information that requires definition in the model) and are passed to the next subsystem level or implementation flow.

Verification of requirements begins with the system modeling and ends up in the system integration. A Verification Capture Point (VCP) is an object that defines how each requirement must be verified, in which design steps the verification activities are to be performed, and what parameters are to be verified. VCP technology enables hierarchical verification that abstracts the verification details to the level needed in every design phase.

Learn more about MBCSE

Interested in digging deeper into MBCSE? Then you can find out more by visiting the Siemens MBCSE Webpage, which offers a next-level dive into this powerful methodology.

If you’d like to take the next step in making MBCSE a reality in your organization, then step up to our MBCSE Adoption Service, which offers more detailed, expert-driven guidance on implementing the Siemens MBCSE methodology and tool flow.

-Susanna Solanti-Iltanen – Digital Transformation Group Siemens EDA