Retrofitting Data-Centricity Fails in Defense M&S

The modernization plan looked reasonable. Take the existing platform-centric system. Add a data layer. Achieve data-centricity without full re-architecture.

It didn't work. It rarely does.

A common pattern in defense M&S: programs try to layer a common data model over service-specific simulation platforms (e.g., one for air, one for ground, one for maritime) without refactoring the underlying interfaces. The result? Another translator layer that maps mismatched assumptions about timing, fidelity, entity resolution, and data quality.

Platform-centric systems encode data assumptions everywhere — in interfaces, storage structures, processing logic, and API designs. Those assumptions aren't isolated in a "data layer" that can be swapped. They're distributed across the codebase, embedded in years of design decisions optimized for a different paradigm. Retrofitting data-centricity into such systems isn't adding a layer. It's rearchitecting through the back door while pretending you're not.

The assumption — that data-centricity can be retrofitted without rework — persists because rearchitecture is expensive and executives prefer incremental approaches. Wrapping legacy systems feels pragmatic. It creates the vocabulary of data-centricity without the structural change.

Here's what actually happens: programs build integration layers that translate between platform assumptions and data-centric interfaces. Those layers — whether "common data models" or yet another translator — become brittle adapters carrying semantic mismatches, timing assumptions, and quality gaps that the underlying platforms were never designed to surface. Under load, they fail. Under change, they break. Under scale, they become the bottleneck.

Data-centricity requires data to be a first-class architectural concern — owned, governed, and managed independently of the platforms that produce or consume it. Systems designed around platforms treat data as a byproduct. No wrapper (or translator) changes that fundamental orientation.

Retrofitting creates the appearance of transformation while preserving the structures that prevent it.

If your data-centric initiative doesn't touch platform architecture, is it data-centric or data-adjacent?