Storage and Space GIS#

OrbitalPV separates scientific history from operational GIS products.

Parquet and DuckDB#

High-volume scientific time series belong in immutable partitioned Parquet datasets. DuckDB is the local analytical query and rollup engine.

PostgreSQL and PostGIS#

PostgreSQL stores shared application state and durable job metadata. PostGIS serves Earth-fixed operational products such as:

  • latest satellite positions;

  • cached ground-track segments;

  • ground stations;

  • coverage footprints;

  • spatial filtering results.

Scientific orbit vectors remain explicitly labelled with frame and epoch.

Frontend level of detail#

Browsers should receive downsampled render products rather than raw high-rate history. Orbit-track caches can be materialized at several sample intervals for selected-satellite, normal interactive, and fleet-level views.

CesiumJS is the primary 3D orbital visualization engine. A future 2D operational GIS view can use MapLibre/vector tiles where that is more efficient.