We are researching different technology for data collection from PLC`s and pushing it forward to AWS.
My questions was meant to understand the performance and limitation in Node-RED.
Sampling will typical be per second.
Is Node-RED meant to be 10-50 tags total or can it handle 1000-2000 tags...
Looking forward to hear from you.
Ethernet comms? yes
Serial Comms?
PLC make and model? Can various but a example is AB CompactLogix Series
PLC protocol you are using? EtherNet/IP
I currenty use Node-Red to gather Weather information from 7 different GPS locations (And 7 different Tags for each) every 3 minutes so that is 23,520 tags read every 24 hrs. and then publish back to my Scada system using Mqtt Sparkplug B protocol.
And on top of that I gather Stock Market Data 5 tags every 3 minute which equals 2400 tags every 24 hrs and this data also get publish back to my Scada Server using Mqtt Sparkplug B protocol.
I run this from a Raspberry Pi 4b that just runs on my VPN.
I have also wrote a program using Pycomm3 that will read tags from several on my Plc's and publish all of then back to my Scada Server using Mqtt Sparkplug B protocol. I have done this for 1000's of tags on a daily basis and had no problems. (However, I'm not currenty using this at the moment)
Just information that might help in your disicion making process.
From my experience node-red is not a bottle neck, main reason is PLC driver optimization.
As mentioned:
But also tag size is important: 16 tags of BOOL type will be equal in size of 1INT. So you should calculate size in bytes of your tag list.
Then comes sampling - 4000bytes of tags from Siemens PLC were read every ~3.5s. While 200bytes took ~200ms. I didn't investigate if dependency is linear.