AiVibeTime-Series Database (TSDB)
A time-series database is a database optimised for storing and querying sequences of time-stamped values, such as sensor readings and machine metrics, with efficient ingestion, compression, retention and time-based aggregation.
Time-series databases organise data by timestamp and series, typically identified by a metric name and tags such as site, machine or sensor. They are designed for high write rates of mostly append-only data, compress values efficiently and provide functions for downsampling, aggregation over time windows, interpolation and gap filling. Retention policies automatically delete or downsample older data to control storage.
In IIoT, time-series databases store data from gateways, MQTT brokers and edge devices for dashboards, condition monitoring, energy analysis and machine learning. Options include purpose-built open-source and commercial databases and time-series extensions to relational databases, and cloud providers offer managed services. They complement or overlap with process historians, which add native industrial connectivity and data quality handling.
Design choices include tag cardinality, since very large numbers of unique series can degrade performance, timestamp precision and time zones, handling of late or out-of-order data, and how quality codes and units are stored. Queries over long periods benefit from pre-computed aggregates. Data governance, backup and access control apply as with any production database, and clocks at data sources must be synchronised for accurate correlation.
Key points
- Optimised for time-stamped, mostly append-only data
- Efficient compression, downsampling and time-window aggregation
- Retention policies control storage growth automatically
- High tag cardinality can degrade performance
- Overlaps with process historians, which add industrial connectivity
Where AiVibe comes in
AiAmbA IoT is AiVibe's edge protocol layer: it discovers devices and normalises industrial signals, with supported protocols including Siemens S7, Mitsubishi MC, Delta Modbus, OPC UA and MQTT, and driver availability confirmed per installation. Writes to machines are deny-by-default.
Related terms
- Store-and-Forward BufferingIndustrial IoT & Edge
- Report by Exception (RBE)Industrial IoT & Edge
- Edge ComputingIndustrial IoT & Edge
- Anomaly DetectionAI & Machine Learning
- Predictive Maintenance (PdM)Quality, Reliability & Maintenance
- Digital TwinIndustrial IoT & Edge