Skellig Automation’s Post

Transcript

The problem though, is when you get into the manufacturing world, I think what people realize quickly is absolutely nobody left to their own devices does anything the exact same way. Like if I asked you to configure what should your will message be, There's like a 0% chance you're writing, Oh no, I died :) right. You're probably going to put maybe the device info, the PLC number or something, I don't know. But The thing is, nobody did it the same. So that's kind of a problem because understanding what is connected to your broker, what is creating hopefully like your process data is a super, super important behavior and thing that needs to be aware of. And it wasn't very interoperable because everybody was doing it differently. So that was one major critique of the MPT specification as it was. The second thing is this connect stuff is totally optional, right? Because with MQTT Explorer we connect to things. All day long and there's no like will message in there because it be obnoxious, right? Like we always use this. Typically I use it right as a diagnostic tool just to connect and disconnect. So that's another problem because somebody could say, Oh yeah, I'm doing MQTT on my PLC, but they also could have configured it wrong and so it goes offline and nobody's any wiser. That's another problem. And then the third kind of big issue was. Again, going back to interoperability. One of the powerful is the double edged sword. With MQTT, you can send literally anything. So again, for your process data, nobody was sending stuff the same format. Maybe people chose Jon, maybe they just did regular text. There was nothing in forcing them to be standard. So it made it hard to kind of build on top of an MQTT infrastructure, at least for process control because you didn't really know what was your data going to look like, right? OK, so for all of those reasons. Eventually they came out with something called Spark plug.