A simple time-series historian designed for sparse or compressed sensor data
NOTE: This service is incomplete and not ready for production use.
Sensistory provides a shard-based, fast recording and fast retrieval time-series database with REST and GRPC APIs.
Unlike many other open-source time-series databases, Sensistory is designed to work well with sparse data
This allows clients to request the condition of a point (or an entire system) at any time, rather than simply at those times when specific sensor recordings were made.
When you perform any query with a time-filter (which will be most queries - Sensistory only works with time-keyed data), the results will include the sample before and after the time frame. This allows you to trim, interpolate or otherwise modify the outlying points to determine the value during the query times even if there were no samples during that query time.
Sensistory can optionally compress your data by avoiding repetitive recording. Due to its sparse data support, the system can ensure that only the first and last records are retained when one of your channels has a stable value for a long time (ie. discrete channels such as enumerations and boolean values). This may save millions of recorded points in channels that only update occasionally.
Sensistory can deliver quasi-live data, delivering the latest values from memory rather than requiring hard-drive access.