Check the
documentation
Ask the
Community
Take a look
at
Academy
Cognite
Status
Page
Contact
Cognite Support
In the Data Foundation-project; there is an understanding that times series data will be represented in accordance with OPC-UA data model using the FDM (DM) functionality in CDF. Is there any issues here in relation to the limitations in FDM stating that there is a limit of 10 million instances? Or will it be designed so that this will not be an issue?
Solved
Scalability issues with representing time series data in accordance with OPC-data model using FDM (Data modelling)
Best answer by Kjetil Halvorsen
OPC-UA can carry various data “types”, and depending on which type of data you are sourcing via OPC-UA, you may want to leverage different resource types / capabilities in CDF.
In simplified terms, we can make the following, high-level guidance:
- Structural entity data. This part of the data typically represents the setup of the control system (from which the OPC-UA data is sourced). You could think of this as the equipment and sensor breakdown structure which helps understand the context of sensors and events. CDF Data Modeling is a good destination for this part of the data.
- Data points, sensor values, variables to be tracked over time. This part of the data matches very well with the CDF Timeseries resource type.
- Events. This part of the data matches well with CDF Events (primarily), or Data Modeling (secondarily).
Correct, Data Modeling (DM) is not intended to store the equivalent of 100s of millions of data points.
You could perhaps use the TimeSeries type in DM and store the data as data points in a Time Series.
OPC-UA can carry various data “types”, and depending on which type of data you are sourcing via OPC-UA, you may want to leverage different resource types / capabilities in CDF.
In simplified terms, we can make the following, high-level guidance:
- Structural entity data. This part of the data typically represents the setup of the control system (from which the OPC-UA data is sourced). You could think of this as the equipment and sensor breakdown structure which helps understand the context of sensors and events. CDF Data Modeling is a good destination for this part of the data.
- Data points, sensor values, variables to be tracked over time. This part of the data matches very well with the CDF Timeseries resource type.
- Events. This part of the data matches well with CDF Events (primarily), or Data Modeling (secondarily).
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.