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 spacecraft become increasingly instrumented with real-time telemetry, operators are facing a critical issue known as the cardinality wall.
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 traditional approach to telemetry infrastructure, relying on relational databases or log analytics platforms, is no longer sufficient.
The cardinality problem arises from the high number of unique telemetry streams, creating a situation where traditional databases struggle to scale. These databases, designed for transactional workloads, face challenges with high-cardinality telemetry. They rely heavily on indexing, which becomes inefficient as the number of dimensions increases. Additionally, they were not built to handle continuous telemetry streams, leading to issues with strict consistency rules and data ingestion.
The issue is further exacerbated by telemetry retention requirements. Satellite programs often store data for years or decades, necessitating both real-time ingestion and large-scale historical analysis. This compounds the cardinality problem, forcing operators to simplify data to maintain operational stability. For instance, Loft Orbital, a microsatellite operator, faced challenges with high-volume telemetry data. By adopting a time series-oriented architecture, they improved real-time monitoring and historical data access.
However, simplifying data comes at a cost. Removing context, such as metadata, can hinder anomaly detection and machine learning systems. Context is crucial for correlating events across subsystems, and its loss can significantly impact mission resilience. As satellite constellations grow in size and autonomy, telemetry infrastructure becomes integral to operational visibility and anomaly response.
Addressing the cardinality wall requires a shift in approach. Operators should identify pressure points, such as delayed anomaly detection or data gaps, to focus on specific areas for improvement. Decoupling the telemetry pipeline into separate components for ingestion, storage, and analytics can help stabilize real-time monitoring. Additionally, preserving full context should be a priority, avoiding short-term fixes that may introduce blind spots later.
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 adapt to this reality. By isolating and redesigning strained architecture parts, teams can make progress. Incremental tuning is no longer a viable strategy; instead, a new generation of LEO infrastructure should be designed with scale, distribution, and context preservation in mind.
In conclusion, the cardinality wall is a critical challenge in modern space operations. It demands a reevaluation of telemetry infrastructure, emphasizing distributed systems, context preservation, and a tailored approach to address specific operational needs. As the industry embraces scale, the impact of telemetry on mission success becomes increasingly evident, making it a pivotal aspect of future space endeavors.
