Bright guy asked me the other day: “Why does any body need a ring topology anyway?” That’s not a bad question. Lots of people don’t understand ring technology and why it’s needed. Yes, It’s pretty much about cost but there’s a few other things to keep in mind.
Most people think that the most important element for choosing a network is speed. If this one is faster than that one then we’ll take the first one. There are a couple of truths here that I’ve been spouting off about for a long time.
BIG TRUTH #1: Nobody selects a network based on any of the network attributes. Yup, that’s true. They don’t pick Modbus TCP because it is user friendly. They don’t pick Profinet IO because it’s fast and they don’t pick EtherNet/IP because CIP is a really cool way to organize a network at various levels of an automation architecture. They pick a network based on what they’ve done in the past and what plc is their favorite. 99% of the time that’s what happens. There’s no closed door meeting to evaluate some huge spreadsheet of features and topologies.
BIG TRUTH #2: Speed Doesn’t matter (at least not very much). In the majority of applications, electrons move over the wire a lot faster than things happen in the mechanical world. There are plenty of things that happen every 2 or 3 seconds. [And as long as I am on big Truths, here’s a bonus big truth for all you single guys. It’s one of my all time BIG TRUTHs: When a woman tells you “I don’t play games…” it means the game has not only started but it’s the second quarter and you’re down by 2 touchdowns!]
I’ll leave women alone for a moment and get back to Ring Topologies… [unfortunately I get paid to talk about networking and not women].
Most of the applications we do are linear. Big paper machine, material handling line or packaging line; that sort of thing. That’s where the most cost savings are for using devices with embedded switches. If you are going to run wires to devices along a 100 ft conveyor line all that wire is expensive as is the installation of it, not to mention all the switches you need. Once those switches are embedded in devices you’ve saved a lot of money but you now have created latency, ways for messages to circle endlessly and a point of failure that can disable every device in a sub net.
If an external switch fails, the switch can be bypassed in one way or another depending on the architecture. Network keeps running. But once, you’ve embedded the switch in a device, if you device fails, no data is passed around the ring and your system is dead in the water.
That’s really what Ring Technology or Ring Topology is all about; managing the performance and failure modes of devices with embedded switches. I’ll have more to say about how CIP Ring technology does that in a day or two.