MBSE Supply Chain Model Exchange: Pitfalls and the Model-Based Design Chain (MBDC)

Source

"Extending MBSE Into Your Supply Chain" — Siemens Teamcenter blog, Mark Sampson (Siemens Product Manager/Evangelist for Systems Engineering and Requirements), November 27, 2019. https://blogs.sw.siemens.com/teamcenter/extending-mbse-into-your-supply-chain/

This is the single most quantitatively specific source found in this research pass, and it directly documents a known MBSE integration pitfall: exchanging architecture models with suppliers is unreliable today.

The Core Statistic

Sampson cites a 2019 Aerospace & Defense PLM Action Group assessment of SysML OEM–supplier data exchange:

Package exchanges with suppliers resulted in successful digital exchange less than 50% of the time, with the ability to interpret design intent correctly succeeding less than 20% of the time.

This is a striking number — it means that even when a SysML package technically transfers between an OEM's and a supplier's tools, the receiving side correctly understands the design intent behind it in fewer than 1 in 5 cases. If accurate and still representative, this is a fundamental interoperability gap, not a minor tooling inconvenience.

Why This Happens (per the article)

• SysML is too flexible. The same solution can be described in many structurally different but semantically equivalent ways, so two organizations' models of "the same" system may not align well enough for automated tools to reconcile them.
• No subsystem isolation. SysML (as of the article's writing) lacked a mechanism to let an OEM share only a bounded subsystem of a model with a supplier — sharing meant either exposing the whole system model (an IP problem) or not sharing a connected model at all (falling back to documents).

The Complexity-Growth Argument

Sampson uses the F-35 as an illustration of unsustainable growth in integration effort:

The F-35 has "175 times the number of SLOC as the F-16" but required "300 times the development effort" — i.e., effort grew even faster than code size, which the article attributes to systems-integration overhead rather than raw feature growth.

This echoes the "systems integration can consume half of program resources" framing used elsewhere in Siemens' MBSE marketing (see mbse-teamcenter-product-positioning.md) — the recurrence across at least two independently-dated blog posts (2019 and undated product page) suggests this is a standing Siemens talking point, though neither source cites a named primary study for the "half of program resources" figure specifically (only the OEM-supplier exchange percentages above are sourced to a named group, the A&D PLM Action Group).

Proposed Fix: Model-Based Design Chain (MBDC)

Siemens' proposed answer, per this article, is MBDC — extending MBSE past a single organization's boundary into the supplier network, built on:

• ISO/IEC/IEEE 15288 standard processes as the shared methodology backbone between OEM and supplier.
• The Obeo/Capella partnership (System Modeling Workbench, launched the prior year — see mbse-capella-system-modeling-workbench.md) as the concrete tooling meant to enable this.
• The goal framed as "start integrated, stay integrated" — i.e., avoid a document handoff step ever existing in the chain, rather than trying to reconcile documents back into models later.

Cross-Reference: Center of Excellence Recommendation

This "start integrated, stay integrated" supply-chain framing pairs with the phased-adoption COE recommendation in the SysML v2 guide (mbse-sysml-v2-adoption-notes.md) — both pieces, six years apart (2019 and 2025), converge on the idea that MBSE success depends on centralized modeling governance/standards ownership, not just picking the right tool.

Honest Note

The 2019 A&D PLM Action Group statistic is the most concrete, non-marketing-sounding data point found in the entire MBSE research pass, but it is second-hand here — Sampson's blog post cites the assessment but does not link to the original A&D PLM Action Group report, and this research pass did not independently locate that primary source. Treat the percentages as "as reported by Siemens, attributed to a named industry group" rather than independently verified. No update or re-run of that assessment (e.g., a 2023/2024 refresh) was found in this pass — it's unclear whether the interoperability gap has since improved with SysML v2 or the various connector products covered elsewhere in this KB.