Telemetry at flight scale, on storage you own.
Arc is the columnar analytical database for aerospace and defense teams who ingest millions of sensor readings per second, query across full mission history, and need their data to stay portable, sovereign, and out of vendor lock-in.
Single Go binary. Open Parquet. Standard SQL.
The shape of the problem
Modern aerospace programs generate telemetry faster than the databases meant to store it can keep up. A single test article produces tens of thousands of channels at kilohertz rates. A constellation, a fleet, or a launch campaign multiplies that across hundreds of vehicles and years of history.
Most teams hit the same wall.
The legacy historian was built before the cloud and licenses by tag. The cloud warehouse ingests in batches and charges egress on every backtest. The time-series database hits a cardinality cliff the moment you tag readings with vehicle ID, mission ID, or test ID. Retention gets cut to control the bill. Detail gets thrown away to survive the volume. And when the post-flight review needs the exact slice of data that explains an anomaly, it is gone.
Retention gets cut to control the bill. Detail gets thrown away to survive the volume. And the data you needed is the data you dropped.
Arc is built for the workload underneath all of this. Ingest the firehose. Keep every channel at full resolution. Query across the whole archive in SQL. Store it as open Parquet on infrastructure you control.
Built for the workloads aerospace actually runs
Vehicle and engine telemetry
Capture every sensor at full resolution. Arc sustains 19.9M records/sec on one instance, with data queryable ~100ms after it lands. Range safety, live monitoring, and anomaly review all run on fresh data.
Ground operations and fleet management
Cross-vehicle visibility for a constellation, a fleet, or a flight test program. Arc handles high cardinality natively, so vehicle ID, mission ID, or article serial number does not collapse query performance.
Post-flight and post-test analysis
Find a signature across a year of tests. Replay an anomalous flight. Compare burns across a campaign. Arc stores history as ZSTD-compressed Parquet, 5-7x smaller than raw, so deep retention costs object-storage prices.
Mission planning and simulation
Historical telemetry is the training data for the next mission. One queryable surface for prior missions, simulation outputs, and live test data, in open Parquet that plugs into Python and notebooks.
What Arc gives an aerospace program
Sovereignty, not lock-in
Apache Parquet on storage you choose: S3, MinIO, GCS, Ceph, or local disk. Open format, every analytical tool can read it. For ITAR, EAR, or sovereign-cloud programs, this matters more than benchmarks.
Single binary, edge to ground
One Go binary, no heavy dependencies. Same database on a ground server, in a sovereign cloud, or on a remote test stand. No cluster to design, no Zookeeper. One operator can run it.
The bill matches the workload, not the vendor
Storage-tier prices for history, not premium-database prices. Compaction and ZSTD cut raw size by 5-7x. Keep full-resolution archives at cold-storage rates.
Standard SQL, no proprietary DSL
Standard SQL through an embedded DuckDB engine. Engineers, analysts, and contractors already know it. No Flux, no q, no language to onboard a new hire into.
Real-time and historical in one place
Streaming ingestion and multi-year archive in the same database. Live dashboards and post-mission analysis query the same Parquet, with the same SQL. No separate hot-and-cold stack.
Why aerospace teams choose Arc over the alternatives
Legacy data historians were built before the cloud and license by tag. Arc is one binary, one license, open Parquet you own. No per-tag fee, no forklift upgrade.
Cloud data warehouses were built for BI, not telemetry. Arc keeps data on storage you own and queries it without egress fees or per-query meter spikes.
Tag-indexed time-series databases hit a cardinality cliff the moment you add vehicle ID or mission ID. Arc stores high-cardinality data as ordinary column data. Fleet-wide queries are routine.
Build-it-yourself DuckDB + Parquet is the most thoughtful alternative, and it is what Arc is. DuckDB plus the 80% you would otherwise build: ingestion, compaction, retention, governance, replication, ops tooling. In one binary.
What this looks like in production
Arc is running in production aerospace flight operations today. Sovereign cloud deployments. Single-instance ingestion at sustained millions of records per second. Multi-year archives in open Parquet on customer-owned object storage. Standard SQL across the full history.
AGPL-3.0 with a commercial license for organizations that need it.
Get started
There are three ways to evaluate Arc for an aerospace workload.
See it run on aerospace data
Our live demos include satellite tracking and telemetry workloads you can query yourself.
Get Arc
Arc is open source and ships as a precompiled single binary. Pull it from the download page as Docker, Helm, or a .deb, and point your existing telemetry pipeline at it. Most engineers have it ingesting within an hour. Source on GitHub.
Talk to us about your program
For programs with ITAR, sovereign cloud, or air-gapped requirements, we run a private discovery to scope deployment, integration, and licensing.