The unit of accountability is misaligned
Program managers are usually accountable for cost, schedule, performance, certification, and readiness of a designated platform or system. No single program naturally owns the complete cross-service mission thread.
A locally rational platform decision can therefore externalize integration cost onto the joint force: proprietary data, closed interfaces, slow release cycles, separate classification rules, and unique test environments.
Platform model versus mission-network model
| Attribute | Standalone platform | Kill-web mission network |
|---|---|---|
| Primary value | Performance and availability of one bounded product | End-to-end mission outcome across heterogeneous systems |
| Requirements | Relatively stable thresholds assigned to one sponsor | Evolving mission hypotheses, interfaces, data semantics, and cross-sponsor outcomes |
| Delivery | Major blocks, lots, and baselines | Frequent software, data, adapter, and policy increments |
| Funding | Development, procurement, and sustainment phases | Persistent product teams spanning build, deploy, operate, secure, and refresh |
| Acceptance | Compliance with a platform specification | Mission-thread performance in representative joint and coalition environments |
| Data rights | Operate and sustain the platform | Interfaces, schemas, test harnesses, build artifacts, adapters, and recompete rights |
| Sustainment | Spares, depots, modifications | Continuous software engineering, vulnerability remediation, API compatibility, and data evolution |
Open architecture is a governance capability
An interface is not practically open merely because one vendor documents it. The government or portfolio authority needs implementable specifications, versioning rules, conformance tests, representative data, technical rights, waiver control, and mission-thread evidence.
MOSA can separate stable physical and safety-critical layers from rapidly evolving mission applications, data services, and adapters. Different layers should have different release cadences and certification paths.
Funding and contracts must support adaptation
- Fund enduring cross-platform products such as identity, data fabrics, adapters, test environments, and integration services.
- Use modular orders and recurring competitions where interfaces and acceptance evidence are common.
- Do not confuse a fast prototype contract with an operational transition plan.
- Price uncertainty honestly and preserve off-ramps, government product ownership, and short acceptance increments.
- Align incentives so vendors can profit from stewardship, integration quality, and repeatable competition rather than lock-in.
Test the mission thread, sustain the software
Every participating component may meet its own specification while the end-to-end path still fails because of latency, semantic mismatch, security labels, unavailable authority, or release drift.
Kill-web acquisition therefore needs persistent joint test environments, automated evidence, authorization reciprocity where justified, digital twins, representative degraded networks, and teams that remain responsible after initial fielding.
Research basis: KW-RPT-005 and KW-RPT-010.