Fusion · aviation data infrastructure

From aviation feeds to usable data.

Fusion is our high-performance catalog and synchronization layer for aviation data, including SWIM-derived feeds. It makes data easier to ingest, update, query, and use across services, applications, and AI systems.

The problem. Every consumer should not need its own downloader, version policy, gap recovery, checksum logic, and local database updater before it can use aviation data.

What Fusion does

One data path for many consumers.

Fusion gives aviation data a durable identity and a repeatable path from published artifact to application query.

Catalogs the data

Fusion publishes version metadata for databases and snapshots, including origin, size, modification time, compression, and checksums.

Moves only what changed

Clients can seed from a snapshot, then apply incremental chunks when the sequence is contiguous. If a gap appears, Fusion can reseed from a newer snapshot.

Makes it queryable

Synced SQLite artifacts back stable query surfaces for operational datasets instead of exposing raw feed mechanics to every caller.

How it works

A catalog, a sync client, and verified local data.

The current implementation favors explicit artifacts and inspectable state. A consumer can understand what it has, what it needs, and why a refresh is safe.

1 · Publish

Feed artifacts and manifests land in object storage. The Fusion catalog describes the current snapshot, retained chunks, sequence bounds, and object metadata.

2 · Decide

A client compares its local sequence and baseline with the catalog. It selects a snapshot or the exact contiguous chunks needed to move forward.

3 · Verify and apply

Downloads are hash-checked, Zstandard assets are expanded as needed, and SQLite updates are applied through temporary files and atomic replacement.

Operational property: a failed hash, missing sequence, or incomplete download stops the update path rather than silently producing a partial database.

Data surfaces

Built for aviation systems that need context.

Fusion currently supports catalog and query paths for a growing set of operational and reference datasets.

Operational feeds

ATIS, RVR, NOTAMs, OOOI, flight plans, TFM advisories, airport demand, and system demand.

Planning data

Routes, callsigns, aircraft profiles, airport and handler data, fuel, procedures, and map-facing navigation artifacts.

Weather and reference

METARs, TAFs, PIREPs, AIRMETs, AFD, terrain-aware layouts, obstacles, and other versioned aviation databases.

Product direction. Fusion is becoming the shared data layer for Aeronautic products and AI operations tools: one place to publish, synchronize, inspect, and serve aviation data without rebuilding the pipeline for every consumer.