2 of 2 people found this helpful
Thanks for reaching out to us on PI Square. There are a lot of different sub-questions in this so I apologize if I miss one of them:
- I've always found the white-boarding video on our Youtube Learning channel to be a good overview of the data flow in the PI System. If you want to get more into the details, all of our documentation is freely available online through LiveLibrary, but the short in sweet version is below
- PI Point Data is updated in AF whenever the data is refreshed on the client. So the most common example is PI System Explorer - anytime you navigate from one element to another or hit the refresh button on the toolbar the data is updated.
- The backend of the Data Archive and the AF Server are separate. The Data Archive is developed on OSI's proprietary archive file format whereas the AF Server is built on top of SQL Server.
- As far as simulated data goes, everyone has their own solution to this problem depending on what data sources you plan on using. Probably the most common one I see if clients using Matrikon's simulation server, though if you're pulling data from SQL Servers perhaps it makes more sense to use something like the Northwind Database from Microsoft. Do you know what kind of Data sources you're going to have? Perhaps we could help identify one specific for your needs.
- Another option if you want a very "turn-key" way to test out the system and it's capabilities is to rent some time on a Virtual Learning Lab VM which already has some faux data set up. Though I should note these are generally set up for specific online courses like the Configuring a Simple PI System Course.
Thanks for the reply. This is useful. One more question:
- The backend of the Data Archive and the AF Server are separate. The Data Archive is developed on OSI's proprietary archive file format whereas the AF Server is built on top of SQL Server. --> How often does the data get pulled into AF Server from Data Archive?
The overview video is having some problem and I cannot access it. I have seen multiple such videos and wanted to reconfirm my understanding. Here is the flow which I have understood - please let me know if I am correct:
If I have a sensor/data source and it registers and sends the stream to the Interface Node (Pi Connector). Since the Data source can send data in various formats, Pi connector will convert these into the format understood by Pi Data Archive and then send the data stream to Pi Data Archive. This will be stored in OSI's proprietary format.
The administrator can setup a element hierarchy of his IoT devices/plant in AF server and map them to Pi Points or time series values stored in Data Archive. So, these streams of data will flow from OSI file (data archive) to AF Server and be stored in SQL. One can also have different types of attributes to elements which are derived, etc.
2 of 2 people found this helpful
Data is not stored in SQL server. All the data coming from interface will be stored on PI Data Archive. Meta data and other relational mappings are stored in AF Server(SQL Database). As your build your hierarchy you add relations and mapping to the PI Points and only meta data is stored. When AF clients requests data it connects to PI Data archive to get data from PI Data Archive server. AF allows you to define your own hierarchy and define your own derived attributes and calculations.
Interface and connectors collect data from data sources and sends to PI system based on instrument tag configuration i.e. item id on source system (dependent on source). PI system process data based on point settings like exception and compression. Data is historised on PI Data Archive. Using AF you can map to PI points/tags and define your own hierarchy based on your requirement and add additional information like location, maintenance date and calculations in AF analysis which will store the meta data in SQL server but point data is coming from PI Data Archive.
Please let me know in case of any queries.
Lal Babu Shaik
Thanks. This is helpful.