Historian Data Categories

There are various ways to store data in Historian Server and there are many operations that can be performed on that data along the way. To provide a consistent framework for data operations in the Historian Server data model, internally the data is divided into several categories and organized into the following hierarchy.

Each category of data has a separate set of characteristics and is handled differently by the historian. Also, some data categories have differences based on whether the data is being handled by storage or classic storage.
 
Original Data vs. Revision Data

Original data is the data received from the data source originally – that is, for the first time. For example, a real-time stream of data from an I/O Server represents original data. The opposite kind of data is called revision data, when you already have some original data stored for some tags and time interval, but would like to correct it by adding some new revision values, and maybe masking some original ones.
 
Original Streamed Data

While the distinction between original and revision data happens from the nature of the data, the distinction between streamed or non-streamed original data depends from the ability of the data source to enforce the time order of its output.
Original streamed data is expected to be in time order where every subsequent VTQ of the same tag has a timestamp not older than the preceding one. An example of the original streamed data could be data received from an I/O Server or DAServer sending a new device reading with the current timestamp every few seconds.
 
 
Real-Time Data

Real-time data is expected to be in time order, where the timestamp is in the past relative to the current Historian Server time (generally within about 30 seconds of the Historian Server system time). Sources of real-time data values are:

• I/O Servers and DAServers sending data with current timestamps
• System tags
• Wonderware Application Server

Late Data

Late data is expected to be in time order, where the timestamp is far in the past compared to the current Historian Server time. Typical sources of late data values are:

• I/O Servers (RTUs) that send data far in the past compared to the Historian Server time
• Application Server that has been configured to send late data to the historian

 
Replication Data

Replication data is data that has been replicated from a Tier-1 historian to a Tier-2 historian.
 
Original Non-Streamed Data

Original non-streamed data can be in any time order. Typical sources of original non-streamed data values are:

• Transact-SQL statements with wwVersion = 'ORIGINAL'
• CSV imports with an operation type of 'ORIGINAL VALUE'
• Fastload .CSV imports
• Original non-streamed data with a timestamp in the future is discarded.

Latest Revision Data

Latest revision data is the addition of new a version of data that has already been stored. Typical sources of latest revision data values are:

• Transact-SQL statements such as INSERTs or UPDATEs with wwVersion = 'LATEST'
• CSV imports with operation types of 'INSERT', 'UPDATE', and 'MULTIPOINT UPDATE'

New-Original Revision Data

The new-original data category is used internally exclusively for replicated data.

Storing Data through HCAL

HCAL can store data from all the data categories:

• Original streamed data, which includes real-time data, late data, and replication data
• Original non-streamed data
• Revision data