A timeseries always contains a time dimension which is materialized by a timestamp.
To qualify and describe a timeseries, you can use optional facets. Each facet is a descriptive information (textual format), which is or can be used for slicing and dicing.
One or multiple measurements are associated to the timeseries. Measurements are the numerical values that can be aggregated (sum, average …) up to provide meaning to your timeseries. A timeseries can contain more than a single measurement, if all the measurements are taken at the exact same moment and always share the same facets.
From a more technical point of view, a timestamp is represented by a date/time type (depending of the underlying database it can be TIMESTAMP or DATETIME), facets by a variable length list of characters (VARCHAR) and the measurements by a numeric value with finite precision (depending of the underlying database it can be DECIMAL or NUMERIC).
