The I-IoT process requires a continuously high data flow from the machines to the central system. This cannot be supported by a standard SQL database. For this reason, with I-IoT we prefer to use technologies inherited from big data, such as NoSQL databases. NoSQL databases are storage systems that are able to manage big data very efficiently by providing a low support of aggregation functions and relationships among data. In the I-IoT, the most common database is the historical database, which can store events, such as alerts or signals and measures, using a time-based method, where the timestamp is the key, or one of the keys, of the item. The most common historical database is the TSDB. Generally speaking, a time series is a series of observations captured over a period of time, but in the I-IoT, it has to support more information. A typical time-series-based item would look as follows:

The NAME column refers to the logic name of the item, the VALUE field stores the measured value, and the QUALITY column, which can be GOOD, BAD, UNKNOWN, or ERROR, stores information about the data acquisition. We can sometimes store additional information such as the attribute and data type.

The following is a list of the most common TSDBs.