DAF CMSO MS Enterprise Governance Is the Hard Problem
The DAF Chief Modeling and Simulation Office oversees a large M&S enterprise with many organizations and major investments. The hard engineering risk is not one model; it is whether shared data, scenarios, and assumptions stay governed.
The Department of the Air Force’s Chief Modeling and Simulation Office sits over an enterprise where governance is the hard problem. In URL-friendly shorthand, the CMSO MS enterprise is not just a collection of models. It is a decision system.
The official CMSO team page says Richard Tempalski serves as the Chief Modeling and Simulation Officer and oversees DAF modeling and simulation investments described as around $3 billion. The DAF CMSO team page also points to enterprise optimization as part of the office’s role.
Air Force reporting has described an M&S enterprise with thousands of employees across roughly 70 organizations. That scale changes the risk. The issue is not whether any single model is useful. The issue is whether the data, assumptions, scenarios, threat representations, and configuration controls stay aligned across the models a decision depends on.
Most public discussion of modeling and simulation focuses on fidelity. Fidelity matters. But at enterprise scale, governance matters just as much. Two models can both be high fidelity and still disagree because they use different terrain, different threat parameters, different aircraft performance data, or different assumptions about weather and sensor behavior.
That disagreement is not always visible. A decision-maker may see two polished outputs and assume the difference reflects uncertainty in the problem. It may instead reflect configuration drift. Without data lineage and scenario control, the organization cannot tell which answer is better or why the answers differ.
The governance work is practical. The enterprise needs authoritative data sources, versioned scenarios, controlled model configuration, documented assumptions, and repeatable V&V. It also needs a way to decide when a local model can diverge from the enterprise baseline and how that divergence is recorded.
The release cadence problem is real. Different organizations update models on different schedules. Some are tied to acquisition programs. Some support training. Some support analysis. Some live in labs with tight controls. Others support exercises. Without enterprise governance, each group can make a reasonable local decision that weakens the shared decision environment.
The data problem should be treated as a product problem. Threat libraries, environmental data, platform performance data, and scenario definitions need owners, release notes, validation status, and consumers who know when something changed. If the shared inputs are unmanaged, model governance becomes a debate after every conflicting result.
Governance also protects speed. When teams know which baseline to use and how to request a change, they spend less time reconciling results after the fact. The enterprise can move faster because it has fewer hidden disagreements.
Polyrhythm focuses on M&S configuration, data lineage, and conformance discipline because that is what trustworthy simulation means at enterprise scale. A model does not become trustworthy because it is complex. It becomes trustworthy when the program can explain its data, assumptions, configuration, validation path, and limits of use.
The DAF CMSO effort is important because it recognizes that modeling and simulation is no longer a set of disconnected tools. It is enterprise infrastructure for training, test, acquisition, and decision support. Governance is not overhead in that environment. It is the mechanism that keeps one model’s answer from quietly undermining another’s.