Schema for Plants, Devices and Signals

Store
, ,
1

Hi, I’m a Student and I am comparing the performance of many databases with time series workloads. I have a question about how to create the best InfluxDB Schema for my usecase.

I have a lot of Plants, and all Plants have one or more Devices and every Device has about 2000 Signals. Every Record of one measurement have the same timestamp (all 10 seconds) and then the 2000 different values (temperature, pressure etc.) I need only very good Read Performance (for example: get the values from 6 signals over 2 weeks), the write performance is not important. What is the best Schema in this Situation? Thanks for your help.

Regards,
Benjamin

2

Hi @jackzampolin

I know this is an old thread, but I hope you remeber what you asked, and more important, how you solved it.
I’m in a similar project, where we have to store metrics of several devices, each with several fields. More concretely, arround 200 devices, with a mean of 30 fields each (two devices may not share any field). The device is identified by three caracateristics (device type, id and location).

My question is, how did you solve your problem? I’m thinking in two approaches:

  • One measuremente per device type: there would be only two tags (id and location) per measurement and each measurement only with device type specific fields (30 fields per measurement).
  • A unique measurement: there would be three tags (type, id and location) and all fields will be in same measurement, some of them with no value (at most 400 fields).

Thanks in advance