The satellite industry is experiencing a revolution, with thousands of spacecraft operating simultaneously in low Earth orbit (LEO). This scale-driven era presents a unique challenge: the sheer volume and complexity of telemetry data. As fleets expand and assets become increasingly instrumented with real-time telemetry, operators are facing a bottleneck that threatens to hinder their operations. This bottleneck, known as the cardinality wall, is a critical issue that demands attention and innovative solutions.
Telemetry, once manageable, has become a distributed systems problem. Modern satellite buses expose tens of thousands of telemetry signals, streaming at sub-second intervals. This results in millions of measurements per second, each with complex structure and metadata. The challenge lies in preserving enough context and fidelity to support autonomous, software-defined fleets at scale. Operators risk turning telemetry into an operational bottleneck if they continue relying on architectures designed for smaller missions.
Ground system databases, traditionally used for transactional workloads, struggle with high-cardinality telemetry. These databases rely heavily on indexing, which becomes inefficient as the number of dimensions increases. They were also built for controlled write workloads, not the continuous telemetry streams of modern constellations. As a result, they can slow down writes, create backlogs, and lead to dropped data, which is unacceptable in spacecraft operations. Moreover, traditional databases are not well-suited for time-based queries, which are essential for aerospace telemetry.
The cardinality problem intensifies when operators attempt to retain telemetry long-term. Satellite programs store data for years or decades, requiring systems to support both real-time ingestion and historical analysis. This compounding pressure forces operators to simplify data, risking the loss of context. Context is crucial for correlating events across subsystems, and it's especially vital for machine learning systems that predict component failures.
Loft Orbital, a company operating microsatellites in LEO, encountered this challenge. As their platform scaled, they needed to handle millions of telemetry measurements per day. By adopting a time series-oriented architecture, they overcame the limitations of relational databases, enabling high-frequency telemetry ingestion, context preservation, and faster access to real-time and historical data. This transformation was essential for reliable LEO operations.
Breaking through the cardinality wall requires a strategic approach. Teams should identify pressure points, such as delayed anomaly detection or data gaps during peak ingest periods. Decoupling parts of the telemetry pipeline and focusing on specific areas of strain can provide a clearer path forward. It's also crucial to revisit strategies that involve discarding data, as these may introduce blind spots during anomaly investigations or system validation.
The solution lies in treating telemetry systems as distributed infrastructure rather than centralized databases. Data arrives out of order and in bursts, requiring systems that can handle this complexity. The key is to isolate and redesign the parts of the architecture that are under strain, rather than attempting a full system overhaul. Incremental tuning is no longer sufficient; the next generation of LEO infrastructure must be designed with scale, distribution, and context preservation in mind.
In conclusion, the cardinality wall is a significant challenge in modern space operations. By recognizing the limits of current approaches, adopting innovative architectures, and prioritizing context preservation, the industry can overcome this bottleneck. The future of LEO operations depends on it, and the time to act is now.
