Hi all,

I'm just traveling home from another industrial PLC4X POC that went extremely 
well. Unfortunately I'm not allowed to mention any names, but I am sure we all 
know the company ;-)

I think this was the first time I've seen PLC4X do some really heavy lifting 
regarding PLC communication. Fortunately the were only some minor glitches in 
the S7 driver, which I managed to fix quickly.

What I really liked, was to get some performance data: on a S7-1500 we 
requested 2600 separate addresses and were able to read these in 190-210ms. 
When polling in 200ms intervals we didn't exceed the I/O limit of 20% on the 
PLC. When we reduced this to 5% of the cycle time, the response times went up a 
little (to about 400ms). But no PLC watchdog complaining about anything.

The load on the client running PLC4X was marginal.

Also we tested the Kafka-Connect plugin and it worked nicely ... We weren't 
able to hit any performance bottlenecks anywhere. Giving a cluster the job to 
collect data and seeing it distribute this load evenly to all the workers makes 
me very confident, that this is a concept that scales well.

Still the plugin needs a little optimization regarding configuration, but I'm 
sure we'll be able to do that very soon.

In the next weeks they want to slowly increase the load by adding more and more 
PLCs and check for unwanted side effects. The final aim would be to connect 
about 1600 PLCs locally ... And about 20k globally ... But let's not get too 
excited and see what time brings.

The more POCs we do, the more I'm deeply sattified with what we have created :-)

Chris



Outlook for Android<https://aka.ms/ghei36> herunterladen

Reply via email to