128. Idea – What to Consider with Industrial Networks Transcript


K. Sorsenginh: 00:00 

First off, Chris I definitely would advise at getting a network assessment because if we don’t know what we have. We don’t know how to improve. Understanding what’s there, existing or what’s there that you don’t know about can definitely move you in the right path of building a secure and robust network. That’s where I would start off at. 

Chris: 00:27 

Welcome to EECO Asks Why a podcast that dives into industrial manufacturing topics, spotlights the heroes that keep America running. I’m your host, Chris Grainger. And on this podcast, we do not cover the latest features and benefits on products that come to market, instead, we focus on advice and insight from the top minds of industry because people and ideas will be how America remains number one in manufacturing in the world. 

All right. So welcome to this episode of EECO Asks Why today we’re going to be digging in with our local expert, Mr. K, who is our solution architect. And we’re going to be talking about some networking items, what to consider with your industrial network. If your network is the ultimate source of information inside your plant, a well-designed one is pretty important but a poor designed one can be a source of inconsistent results, or even at the worst outages. 

So most people are familiar with the office of Ethernet, but we’re not talking about office Ethernet today. We’re gonna be talking about your industrial Ethernet. One place to start is with assessing your network requirements. So K , you’ve been out there a long time. You’ve seen a lot of different things. Can you give us some of what you’ve seen with your experience when you’re assessing a plant’s industrial network? 

K. Sorsenginh: 01:51 

Yeah. So typically when we head out to plants a lot of items are not documented like they are in the IT-enterprise world. So we start off by cataloging devices. And just noting where all the components are, the connection points, that’s the initial process of logging the IT equipment that are utilized within the plant floor. 

Chris: 02:16 

Okay. So just taking a good catalog of what’s there. Okay. So if you look at like a… Give us some examples of where you walked in and you had that, what in the world is going on here experience, but what does that look like? 

K. Sorsenginh: 02:32 

Well, typically in the OT side of the world, production drives the movement of everything. When there’s a need of additional ports, people are running to Walmart or Best Buy grabbed the first switch they see and throw it in there, right? 

When they need wireless access points, they may grab something very similar, put it on a wood platform and just hang it off of the ceiling or just wrap it or tie it up into the column. Without real protection or any consideration at all. We’ve been into plants where they’re running multi-billion dollar processes and we have hubs laying on the ground covering two inches of dust. 

Chris: 03:21 

So that’s, that’s not good. 

K. Sorsenginh: 03:23 Not at all, not at all. 

Chris: 03:26 

I mean we’ve all seen it been in the plants and you walk into that network. A lot of times it’s a closet in the way it’s treated and it’s just amazing, so much important data and how it’s treated from that structure. So you mentioned components, maybe let’s start there with our listeners. If you’re really trying to consider your industrial network and how to make it robust and reliable and efficient, components are a great place to start. The first thing you think of, when I walk in the room, I see the blue hose. So what should listeners consider when they’re looking at their cabling?

K. Sorsenginh: 04:01 

So with the physical medium, all determines through the MICE environment, right? So you got mechanical ingress and so on and so forth that you would associate the type of cable-in or medium that are utilized.

One of the other things that we have to look at is the environment within that facility right? So if there’s a lot of. Water, chemicals and any other types of substances that would affect, the cable-in and we got to consider what type of jacketing we use. Is it a shielded noise environment?

We typically would run through plants and see items such as Ethernet cables running parallel or in the same wire conduit as 480 higher voltage stuff. Sometimes you cannot get around it and you do have to run it in there, but there is also cabling specified for those like 600 volts high shielded. So selecting the right media for transmission is a definite must in determining the proper build of your industrial network. 

Chris: 05:13 

We hear Cat5 and Cat6 a lot now. And I’ve heard you several times really encouraging, listeners and our users to, to consider Cat6. Why make that choice? 

K. Sorsenginh: 05:25 

Well, because the typical old infrastructure out in the field now you’ll see down to like a Cat3. These were the original Ethernet. So let’s take a step back and look at Ethernet itself. It’s when we’re looking at Ethernet, it’s a, I triple E standard, right? That basically on 802.3 is that your standard Ethernet, and then your 802.11, is your WAN which is the wireless. 

Chris: 06:00 

So WAN is wireless? 

K. Sorsenginh: 06:02 

Correct. And your LAN is basically the hard connection. All these are the typical same Ethernet standard that your IT department uses. And everywhere else. So within the industrial environment that has been adopted from your IT side. So it is more acceptable to utilize an Ethernet out in the production floor because they’re familiar with it. With that being said back in the days, Cat3 were the latest and greatest, and those were installed well with typically newer components that utilize Ethernet, the bandwidth are now being exceeded.

So we keep adding more components and it exceeds the capability of the cable. And we went with Cat5, which was an increase in frequency. And even Cat5 we’ve reached the limit, we’re pushing boundaries. Things are getting faster. Processors are getting faster. If you think about it, it’s just like everything else, right?

Your watch has more processing power than the challenger that shot out in the space. It’s crazy if you think about it. Movement with the, industrial 4.0 IIOT, all this has been pushed down to utilize information coming out of components. So anything from PLC down to sensory level, we’re getting information and just pushing it in on, not even considering what’s there as the infrastructure. So as people started to experience these type of connectivity issues. They started increasing the raps, the frequency. With the latest standard they were able to do up to, like a Cat5 E 400 megahertz or 500 megahertz, which, you could even run, gigabit speeds now off of those. 

But even those now are getting to the point of meeting or getting close to the bandwidth. So as industrial standard comes out, they start increasing the shielding. If you cut open a Cat5 and a cat5e and a cat5, you’ll see the cable is a little bit thicker instead of using the 28 gauge they’re now using 24 or 22. The twists in the pairs are a little bit tighter to prevent the crosstalk, right? There’s shielding, there’s plastic barriers in there. You’ll notice the Cat6 are a little bit thicker and that’s why.

 If you look at the 600-volt Ethernet cable that I mentioned earlier. You’ll see, it’s probably like three sides of a standard straw versus your standard Ethernet cable. And the reason for that it’s got a lot of shielding and it’s got a lot of installation to account for those types of noise or protection. But yeah, Cat6 has been out there a while.

And, we’re seeing Cat7 and even now with our fluke DSX- 8000 it’s capable of testing Cat8 which I went to a customer yesterday and showed him my meter and he says, “Wow, this is Cat8?” This is an IT guy. And he was like, “Wow, are you seeing this?” I said, “Not a whole lot, but this is the new standard coming out.”

So yeah. Yeah. I said, “Think about it. You’ve got some Cat5s and that made Cat5 easier. We’re already looking at Cat8s. That just kind of shows you how far your system is behind.” And it was a shocker, even though he’s been in that industry for a while, seen the new capabilities of cabling and infrastructure upgrades. It’s a sight to see. 

Chris: 09:50 

I bet. And it’s with data these days, right? Everybody’s moving more data, they’re pushing more data, they want to get it, resources to make better decisions on the plant floor. So that media is really important. So thank you for walking through that.

So you lead the cable and then you go into a, some type of hub or a switch, right? So maybe, can you give a little explanation of what you’ve seen, what they should consider when they’re looking at hubs and switches as well from an industrial standpoint? 

K. Sorsenginh: 10:20 

Yeah. And, I mentioned earlier facilities having these switches that you can purchase from your next-door retailer. These are not designed for your industrial environment. When we’re looking at industrial Ethernet, we should consider not only the cable, we should consider the hardware. Is it built to withstand the heat? Typical enterprise-level switch. You know, they’re great for doing emails and sitting out in server rooms, but when you get out on the floor that ambient temperature, may rise as high as 104 degrees, 40 Cs, 50 Cs, whatever the case may be. And those switches and hubs are not designed for that type of environment.

 You know, taking a step back when you’re grabbing hubs, which you’re seeing less and less of nowadays, but they’re still available looking at a hub. All it does is repeat that information throughout every single port, right? So you’re just creating a big old blast of information to every single device that doesn’t even need it. 

Well with switches. It’s improved that technology it’s able to distinguish. This guy doesn’t really need this information, so I’ll just send it to him. And this all deals with Mac, right? Your layer two, where it’s okay I’ve got this Mac down, this Mac table correlates to this machine, so it kind of gets better.

But in the industrial sector, switches that are non-industrial rated versus industrial rated. Other types of protocols or I would say functions are enabled from the factory. Things as IGMP snooping, right? For Ethernet, IP, or querying snoop and so on and so forth. Different Ethernet devices that are designed for the industrial are designed for specific industrial protocols.

And we should definitely consider industrial switches and hardware due to the fact that sometimes you can’t put it just directly in a nice little cabinet. The connection points with water, moisture, being available we got to look at the connectors, right? We got to look at, is that IP65 rated? It is 67 rated? 

Even some of the more robust where there is a lot of vibration, we have to consider the M12 connections. And these basically screw on versus your standard RJ45 plugs because typically, if there’s a lot of vibration, it may get to the point where it may vibrate out with the M12 and the decoded connection point they’re screwed in, and they’re pretty robust and solid to where they won’t vibrate out. The ingression level is a lot higher, so switches that are designed to be IP65, 67, we’ll tend to utilize M12 and the D-Sub connectors for the ethernet connections. 

Chris: 13:27 

Okay. So thank you for walking us through a lot of those components that really make that robust industrial network right? And I guess the point of all of that is we want a network that performs well. You know, when you think about performance considerations, so one thing that we’ve seen over the years is a lot of latency and jitters. Can you explain what that experience would be like for an end-user on the plant floor? When they’re seeing that jitter latency through the industrial network. 

K. Sorsenginh: 13:58 

So if you look at the term latency, it’s basically the processing of information from one node to the other. Jitter on the other hand is the difference within that same timeframe, there’s inconsistency. So in the industrial environment, if you’re looking to do some processing this doesn’t entail sending emails out because if I get it in two seconds, versus if I get in five, it doesn’t really matter.

But in the industrial plant floor, there’s things such as motion safeties, right? So when you have these jitters and latency, these packages are not arriving at a specific time. They’re just everywhere. And especially with motion, if you’re looking to cut a piece of product. 

For instance, if we look at a panel and we’re trying to cut a four-inch panel, at a thousand feet a minute, if we can’t process that information quick enough, We’re going to have some issues of being short, being long. It’s difficult to maintain standards if we can’t have the information there at the time that we need. So that’s basically in a hard shelf of latency and jitters. 

Chris: 15:22 

Okay. So other things from a performance consideration standpoint outside of the media, as we talked about is to topology, right? Of a network. So maybe you can just high-level, walk the listeners through from a topology standpoint, the things to consider with your industrial network. 

K. Sorsenginh: 15:42 

So on a typical production floor, we should expect to see a backbone that could either be comprised of copper or fiber, right? Obviously on the fiber aspect if it’s under, two kilometers we’d be using multi-mode.

For instance, plants that are a couple miles apart, and then we would use a single mode of fiber connection between plants. That being said, once the backbone is there if there is even a backbone typically it distributes out to an IDF panel, which is out on the floor. IDF is basically like an intermediate connection to the MDF, which is your main distribution frame, which resides in like a server room.

 From there on out that’s the typical hard stop line for your IT department. Once you get into the LT side, we have to look at, what are we gonna use? Are we gonna use a standard, a home run connection, which is your star topology? Are we going to use a linear connection, which is basically a Daisy chain from one device to the other?

Or we’re going to look at a mesh or are we going to look at a ring? So these are all types of topology that need to be considered. And again if we’re going back referring to the retailer’s switch and, guys randomly connecting there’s really no topology. There may be a mixture.

You may see some issue when a guy connects a port, runs around another switch and comes back to that same, signalized hardware. You’re going to create a loop, which is going to broadcast storm everything. And going back when we’re talking about switches. Industrials are smart enough to detect these types of loops and would disable things such as rapid spanning tree.

That’s what they’re designed for to mitigate these loops and shutting down machines and so on and so forth. So when the typical connection from a cell level, right, a machine. It’s isolated to that particular machine there’s a primary switch that’s involved and everything else will typically be ran to a cabinet.

Well, if there’s a lot of components within there and they may go from a star topology, out to the next switch and do a linear topology. So depending on the application, the required speed and of course the budget it will determine what type of topology you select. 

Chris: 18:21 

That’s very good because that was my follow-up question. It was okay is there a guideline, but it sounds like it’s application-specific budget constraints. All those things come into play when you’re trying to make those decisions. 

K. Sorsenginh: 18:32 

Speed is also a key factor to this. And as I mentioned, if we’re using this particular topology for a motion application, we’d want to throw as much money and performs that as possible. So that you’re able to grow from there. Be a little bit modular. 

Chris: 18:50 

Absolutely. Absolutely. We’ve covered components. We’ve covered performance considerations. Let’s finish up with talking about assurance and security. That’s coming up a lot more and more when we’re talking about networks and security and cybersecurity. We hear the terms, confidentiality, integrity, availability when people are talking to us about assurance and security of their networks we see availability, integrity as being really two key components that users are concerned with. What’s your experience with this field and what should users consider when talking about items like security with their industrial networks? 

K. Sorsenginh: 19:29 

So based on standards its best practice is to utilize every single component that’s available to you. So if you have switches that are out in the field especially in the industry we would typically recommend you utilize a managed switch, right? A managed switch allows you to do control access. It allows you to control which ports are enabled or which ports are disabled. A lot of times the intrusion comes in from internally, right?

Your third shift guy brings in his laptops and wants to connect to, I don’t know, LimeWire or whatever they’re using nowadays. And that’s how things get infected. Even at the physical level, unused ports should be capped off. There are plugs from Panduit, from whoever that will physically lock you out from that port.

So that’s the physical security, right? If we mitigate user from accessing it, and that’s the first step. So as with anything else in guidelines, CP, WE, or whatnot. We should look at the castle approach of security where things overlay each other. So we block out ports, we disabled ports. You know, have policies in place. Industrial switches will have these types of capabilities to lock people out.

 When we’re looking at switches and credentials. Every switch has a default credential, right? Those that are commissioning and implementing should take the necessary steps to change your password so that they’re not easily accessible. It sees little simple things that will handle 80% of your intrusion. 

Chris: 21:25 

So Password is not password or 1, 2, 3, 4?

K. Sorsenginh: 21:28 

Correct. That reminds me. I have to go change mine. So.

Chris: 21:32 

Yeah, go change that password, K. You know, one thing you mentioned just then with CPWE and converge plant-wide Ethernet, can you talk to that just a little bit? Cause one of the questions I had was about resources for our listeners. If they want to learn more to put themselves in a better situation with their industrial network to CPWE may be a good guide or a framework for them. And I know you have a lot of experience in that world. 

K. Sorsenginh: 21:58 

Yeah. So CPW was basically designed and written now by Cisco. Rockwell Automation utilizes it quite a bit. So there’s a couple of partners that are, that really rely on this Panduit. I commonly refer it to the network Bible. You should definitely take a look at this, in conjunction to TIA 1005, your IEEE 8000 or 802.3.11 to get familiarized with industrial standards. 

But yeah, that guide will walk you through the converge plant-wide networks. As we talked earlier, we have information coming from every single sensor. It gets to the point where it gets to the processor. The processor now pushes it up right into your next level, which then would be like a DCS system or something.

From there on out, it may go into an ERP or MES system. And eventually now maybe corporate we’ll say well, I want that information. How do I get it there? CPWE converges that OT and IT side. So we’re able to pass information through, but still maintain the security and the integrity and the availability to your industrial network and components.

Chris: 23:19 

Very good. Well K, you have a lot of experience in this field. I’m sure a lot of our listeners would just like to know if there are one or two things that they could do, low-hanging fruit, to improve their network. What would those pieces of advice be from you? 

K. Sorsenginh: 23:38 

First off, Chris I definitely would advise at getting a network assessment because if we don’t know what we have. We don’t know how to improve. Understanding what’s there, existing or what’s there that you don’t know about can definitely move you in the right path of building a secure and robust network. That’s where I would start off at. 

Chris: 24:06 

That’s great advice. Great advice, because you really need to know where you’re at, to know where you can go. So, Kay, thank you so much. We hope that the topics we covered today. We’ve covered components, performance consideration, security. There’ve been a lot of good tidbits in here. Please continue to follow us and thank you for your time, K, today on this topic.

K. Sorsenginh: 24:29 

Thanks Chris.