A federated cloud-edge mission mesh
Enterprise services are suitable for durable catalogs, historical data, global planning, software distribution, and large-scale analysis. Theater services can coordinate regional fusion and orchestration. Tactical and platform edge nodes must preserve time-sensitive sensing, policy enforcement, local control, and disconnected operation.
The placement rule is simple: put a function at the slowest and least exposed tier that still meets its deadline and resilience requirement.
Seven interacting layers
1. Sensing and sources
Air, land, maritime, space, cyber, partner, and intelligence sources with calibration, time, confidence, and provenance.
2. Edge processing
Filtering, detection, correlation, compression, caching, and local fallback near the source.
3. Data fabric
Discovery, schema mediation, lineage, access, synchronization, storage, and conflict handling.
4. Transport
Diverse bearers, routing, priority, store-and-forward, quality of service, and path health.
5. Trust and policy
Identity, device and workload posture, data labels, releasability, safety, and authority.
6. Orchestration and C2
Capability discovery, constraint solving, reservations, recommendations, command decisions, and tasking.
7. Effects and assessment
Authorized abstract effects, status, outcome evidence, accountability, and feedback.
Semantic interoperability
Transporting a message is not the same as understanding it. A usable operational object needs identity, observation time, coordinate reference, uncertainty, source, transformations, quality, classification, releasability, validity period, and schema version.
Gateways and generated mediation can connect legacy formats, but translation must preserve meaning, timing, trust context, and failure behavior—not only syntax.
Cloud-edge placement and disconnected operation
| Tier | Best suited to | Failure rule |
|---|---|---|
| Platform or local edge | Immediate control, filtering, local tracking, safety, and bounded fallback | Must not require a distant synchronous service |
| Tactical edge | Local fusion, mission caching, short-range orchestration, delegated execution | Must operate through intermittent transport and reconcile later |
| Theater or regional | Cross-unit fusion, broader resource composition, replicated mission services | Survive loss of one region or gateway |
| Enterprise | Global catalogs, history, model development, software distribution, deliberate planning | May enrich or supervise but should not close the fastest loops |
Mission-specific latency, not one marketing number
No public universal kill-web latency threshold applies to every function. Local control may require milliseconds; track exchange may require subsecond or seconds; human decision support may tolerate longer; cross-domain resource recomposition of the ACK type was publicly described on the order of minutes.
Requirements should specify start and stop events, percentile, deadline-miss rate, data age, jitter, load, security processing, and degraded conditions. Average latency alone is not an adequate acceptance criterion.
Resilience metrics that matter
- Number of independent, authorized, mission-capable paths
- Time to detect a failed or untrustworthy node
- Time to validate and commit an alternate path
- Mission completion under defined correlated failures
- Maximum safe disconnected operating period
- Data freshness and deadline-delivery probability
- Ability to reconstruct the path, evidence, policy checks, and human decisions
Research basis: KW-RPT-002, KW-RPT-008, KW-RPT-009, and KW-RPT-011.