Ep 06: IIoT extends control of your Operations to the very edge

00:04
Speaker 1
Hello and welcome to the Human and Machine podcast. The Human and Machine podcast is, of course, targeted at our south African audience. We based here in Johannesburg in South Africa. My name is Jaku. I’m with my co-host, Lenny. And we are talking about, in every episode, we are talking about anything and everything. Industrial automation, industrial technology, manufacturing, and the world of system integration and all of the new advances and all of the new observations around what is happening in this industry that we love. If you missed last week’s episode, sure, Lenny. It was a cracker.

00:39
Speaker 2
Yeah, it was a very, very valuable topic. Very insightful from Arlen, co-inventor of the MQTT protocol. Just the story and the way that the guy tells the journey of how MQTT came into the market, the reasoning about it, the applications that it can help, especially in the IoT kind of space. It was a very valuable topic, and again, very humble for Arlin to join us, actually on that session.

01:08
Speaker 1
Absolutely. And what an absolute legend of the industry. 42 years, MQtT, genesis in the early eighties. Hard to believe. Hard to believe it’s been around for that long. So that was a really insightful episode with Arlen Nipper, the co-inventor of MQTT. And probably a nice segue into this episode. Our guest this week that we’re very much looking forward to chatting today and more of a local flavour around this world of IIot is Louis van Weke. So Louis joins us from business Connection. Louis is the managing executive for IoT and OT solutions. And again, it’s a nice little segue from last week’s discussion around MQTT and what that looks like in South Africa, definitely.

01:53
Speaker 2
And I think going a little away from the pure technology that we had with holland, a little bit more into practical implementations and what we’ve seen in the African market. I know a few years ago, Africa was pointed to be one of the biggest places where IoT would be revelants. And kind of to understand from his experience, why haven’t we seen this massive uptake in the buzzword of Iot in the country.

02:21
Speaker 1
Absolutely. So, Louis going to share wealth of his experience in the South Africa market. And Louis, that was your introduction. It was a long one, but that was your introduction. Welcome and thank you for joining us.

02:32
Speaker 3
Yeah, thanks, Jakub, Lenny, thanks for having me on human machine. Very excited to speak to you guys today. I think we’ve got an interesting topic, something which is very current in the industry and possibly top of mind for a lot of people out there. So, yes, I’m very happy to be with you here today.

02:53
Speaker 1
Cool. Fantastic. Louis, we’ve known each other for quite a few years. We were actually chatting a few weeks ago. We were chatting with Francois. I think the last time that I saw Francois was in Sun City. And I think the last time that I saw Louis was also in Sun City. And we made the joke a few weeks ago. Unfortunately, we’re not in Sun City now. The world is very, very different today. And probably given the space that we’re in and the limitations that we’re operating with at the moment, probably for IoT and more importantly, IoT, some really good opportunities. But maybe before we get into that, if you don’t mind telling us a little bit about yourself, you’re definitely not a junior in the industry. You’ve been around for. For quite a while. Where did your journey start?

03:42
Speaker 3
Yeah, I’d say I’ve been quite a traveller with regards to the industry. I’ve not always been in the field of engineering, automation and industry, but I’ve very much started off there by training. I’m an engineer, chemical engineer. So please excuse me if I’m at points become a bit analytical. I think that comes with the territory and that sort of background. We like to analyse everything we see. So that was my career start. And initially I spent about ten years in the industry, predominantly in steel making and in mining various different plants and facilities. And then I took a complete left turn in my career and I went into management consulting, which had much more of business flavour to it.

04:43
Speaker 3
And I spent quite a lot of years in that before I moved into the IP telecoms environment with BCX and the telecom group, where I’m now in that environment. I’ve also been involved in a number of businesses which is more related to the pure it and not so much the industry things such as software development and some it consulting and so on. And then the opportunity came up to for me to actually hit up the IoT and OT solutions business within the BCX group where I am now, which sort of took me back after a detour of about ten years, almost as well back into the. Back into the industry and the IoT and operations technology space where I am now.

05:38
Speaker 1
I’m keen to know if after ten years, have you found that a lot has changed or not when you returned? Probably not as much. As many things change, it stays the same?

05:47
Speaker 3
Surprisingly, yes, that is probably the answer. I think a lot of things change on face value, but deep-down underneath, you know, you realize that it’s to some extent a state the same. And I still felt very comfortable, you know, getting back into the industry, possibly because of this, that it hasn’t changed so fundamentally. At least the thinking and the drivers in the industry, I think, remain constant. And that really set the scene.

06:19
Speaker 1
Absolutely. I think the drivers and the fact that you mentioned the business aspect, I mean, it’s something that we’ve been hearing as a recurring theme on the episodes is about the business case and the business need. And I think probably around Iot or specifically IIot. And maybe that’s a good departure point, is for some of the listeners that are wondering exactly what the difference is between IoT and IoT, we’d like to, again, we’ll start there, but maybe then also just understand Iiota, what are some of the drivers that we’re seeing and some of the practical use case applications? I mean, there’s a lot said about condition monitoring, predictive maintenance. I mean, there’s so many different areas. But maybe let’s kick off with IoT versus IIot for the lame person out there.

07:11
Speaker 3
Yeah. So I think that’s a bit of a big question you’re posing to me because personally, I do not see a very big difference between the two. Obviously, the naming convention is coining a term to focus IoT specifically on the industrial side of things, calling it iiot. But on a practical level, the one difference I think that we may experience out there in the industry is possibly the underlying technology in the environment that you’re dealing with. If you’re dealing with industrial IoT, you’re typically looking at manufacturing operations, mining operations, and as you know, these plants and mines typically have mature control systems monitoring operations, which means Iota in the industrial sense, often involves integrating into those systems.

08:16
Speaker 3
Whereas I think IoT, in the general sense of the word, would be more separate technology installed into an environment with less integration and less having to deal with the industrial stuff, such as the control systems, the PlC’s, and so on.

08:39
Speaker 1
And.

08:39
Speaker 3
Yeah, but there’s exactly this. There’s pretty much overlap in everything. So if you look at something like, for example, building management, it doesn’t look that different because you’re still integrating into PlC’s, controlling some parts of the buildings, doors, the H vac, etcetera. So from that point of view, it’s not that different. So, personally, I would say, yeah, for me, the two topics are very closely related, and I think it’s really just a way of distinguishing what sort of industry or vertical we’re dealing with. If you’re talking about industrial Iot, versus general IoT.

09:26
Speaker 1
Yeah, and something we mentioned a few weeks ago, Lenny, is that often in the industrial world, a lot of the equipment assets, things that you want to connect and put into the IoT space, very often those things were made not to be connected and not to be put into this world and built and engineered and placed down to last a lifetime. And I suppose that’s where the conversation of, when we’re talking about digital, which this is, how digital is it out there, Louis, and how analogue is it still? Maybe that’s the reality of the industrial world, and maybe that’s something that you come across every day.

10:10
Speaker 3
Yes. Jaqi. So from my point of view, I think we find that there is, we do not really have such a big problem with the difference between analogue and digital, but that’s mainly because of the nature of the business that we’re running, which is a combination of working in the operations technology space as well as the IoT space. So in terms of skills, you know, I’ve got a lot of people working for me that’s very capable, you know, in both environments. But down to your question.

10:47
Speaker 1
Yes.

10:48
Speaker 3
In the industrial space, there is still a lot of technology deployed out there which is analogous, older technology, but this technology has always been implemented to be able to communicate, you know, with control systems, etc. That is, you know, on the plant level, the big difference being that this, these technologies were never designed, you know, to be Internet connected.

11:17
Speaker 1
Yes.

11:19
Speaker 3
And, and that is, for me, perhaps the big distinguishing factor. And what has changed in terms of Iot, a lot of the work we do in the IIot space has to do with putting in devices, software, middleware, etcetera, to solve that communication problem, from plan to central server to cloud.

11:41
Speaker 1
Yes.

11:43
Speaker 3
And sometimes you need to actually have physical devices in addition to what is already available that you need to integrate with to achieve that. So I would say that’s, for me, the dominant difference between the two. But you’re very correct in saying that there’s still a lot of technology out there in the field, specifically in the manufacturing environment, which is not, in principle, designed for the Iot world, but it definitely has the capability, if correctly integrated.

12:14
Speaker 1
Yes.

12:16
Speaker 2
And Louis, maybe taking on that topic, I mean, we’re talking about industrial Iot, where, you know, we’re lucky in that sense that a lot of our machines, a lot of our pieces of equipment are controlled. So there is a sense of a digital face where we can actually interface to and make that then available to be pushed into a cloud kind of scenario when we now talk about just normal Iot, where, you know, there is not a sensor, there’s nothing there. You need to, I mean, we see these promotional videos where a guy takes an id plaque, puts it onto his pot plant, and now it’s sending the live. But I don’t think that’s really the case as those promotional videos make it so easy and seamlessly to do.

12:59
Speaker 2
I don’t know and I don’t see that we’ve seen that really massive uptake of all of these massive sensors and device connectivity at this point time.

13:07
Speaker 1
Well, yeah. Louis, have you seen a massive uptake? I think, you know, we’ve seen varying reports. Yeah.

13:15
Speaker 3
Yeah. So I think we’re in the field here of a bit of mixed emotion because personally the uptake I’ve seen, you know, in terms of the hype in the market is huge. Yes, physically, in terms of what you really see out there, in terms of solutions, it’s much more limited. But there are certain solutions which has become sectors in the market by themselves, which on its own would be big, but there’s big barriers for entry of getting into that. And typically, those type of solutions has been around before, traditionally, and has been sort of rebranded into IoT. It’s maybe not as pure as people would think, but if you count those numbers, it adds up to some of the big numbers you see in IoT. But that is not a reality in terms of the new business that’s available out there.

14:10
Speaker 3
But to get back to Lenny, your original point, so how easy is it to just get the world out there connected? I think that is really the key behind all of this. So, we all understand the industrial space fairly well. We know there’s a lot of equipment, sensors, et cetera, out there in most of the manufacturing facilities. And the real question is, how do I integrate and connect? Our move today is more into other environments where we implement pure IoT, such as a retail environment, and the integration requirements in those environments becomes entirely different. Let’s think about some practical examples. Let’s look at a retail store. You know, what would be the type of things that, of importance in that sort of industry? One simple example would be, for instance, their cold chain and their fridges inside their stores where they’ve got perishable goods.

15:20
Speaker 3
Many of these physical equipment that I’ve got stores doesn’t have inherent capability, which means you need to install sensors in there. And even though it looks similar to the industrial environments, in other words, that it’s brownfields, and you were dealing with existing install dice, the solution to the problems are typically very different. You would have to install the entire chain from sensor to gateway to platform, with everything in between, you know, with the comms on a local level, comms to cloud, etcetera. And the drivers in the, in those industries are also different. Cost, for example, being one that’s, that’s a big, big problem, you know, for to get massive uptake in IoT.

16:21
Speaker 3
If you look at the industrial environment, I think the decision makers in that space are more comfortable because the equipment that used to be installed there are industrial grade, reliable, robust, you know, and typically those things comes at a higher price point. So there’s a bit of an easier acceptance, you know, on that level, for the price point. If you look at retail, they literally calculate the cost of what they spend on all technology. Everything in store gets discounted into the price of the products that they sell, and literally down to the percentage it makes up of the price of the product it’s sold. So, there’s a very strong driver in terms of cost, which pushes IoT in the direction where it needs to look at cheaper and simpler ways for solving those problems.

17:24
Speaker 3
But you’re still sitting with the same type of basic issues. It costs money for the labour to actually install it. And if you look at something like retail, the big attractiveness there would be, economy of scale. It is the number of installations that you do that actually makes it attractive. But the more installations you need to do, the more workforce you have to actually cover that. So it creates all sorts of new business challenges, which we didn’t really have when we’re dealing purely in the operations or industrial space, where you typically had engineers dedicated to a certain plant or plant areas or particular clients, and they could cope with that installation and integration needs that you have. And now you may have to roll out the installation for, let’s say, a big retail chain for no specific reason.

18:23
Speaker 3
Let’s just choose something like Shoprite checkers. If you think about them, they’ve got thousands of stores spread out all over the country. And that’s a completely different challenge in terms of the physical field force that’s required to get in store. You need to have small, repetitive solutions, easy to install, easy to configure, easy to manage remotely. Communications becomes a big issue because you need to think about how would the communication work in different localities. You don’t have just one place, and let’s say I’ve got fibre connected here, problem solved. I just make everything communicate over that one, you actually need to solve the problem more generally so that it would fit all those different places or point of installations that you need to deal with.

19:21
Speaker 3
So, yeah, I think the complexity there in terms of trying to find a general solution to a specific problem makes it difficult. Whereas if you look at the industrial environment, it’s down to hard tech, you know, thinking about things like your typical communication protocols, you know, things like Modbus, profinet, Dacnet, etcetera. But we know that world, you know, it’s always been the world we worked in. So I think we’ve got a bit of a comfort zone there. Maybe we’ve got a bit of a blind spot, you know, when we get to things like retail banking and so on. Sure.

20:03
Speaker 2
And Louis, I think you made a valid point there that if we look at, you know, the typical kind of IoT solutions that we need to provide, you need to provide a sensor, you need to provide the local network to connect to that sensor, you need now to make that available to be, you know, pushed to whatever network to be actually now stored in a cloud solution where you’re going to do whatever business need you need to do. On that topic, but I almost feel sometimes that the perception out there is that this sensor, this little local network that you now need to enable on this site, I think the perception is that it is supposed to be cheap. I mean, the sensor should be a few hundred rands because of the scale of what you need to install. Is that the case?

20:46
Speaker 2
Is sensors really that cheap to actually get going? Or if you do that quote for the guy who looks at the price and like, what are you trying to sell me here? But I just have this feeling that we’re not there yet with sensors being to that point.

20:59
Speaker 1
Well, and more importantly, how does that scale?

21:01
Speaker 2
How does it scale?

21:02
Speaker 1
And maybe to add to that, Lenny, and a question for you, Louis. I love the example of the cold chain that you mentioned. So on the topic of high upfront costs, Lenny mentioned the cost of the devices. I know there was a survey by Frost and Sullivan that talks about 60% of manufacturers worried about the costs associated with developing IIot and getting things connected. So to your example about the cold chain and these devices, the fridge, the industrial fridge that doesn’t ship with anything connected yet, at least not as far as I’m aware of. Maybe that’s fairly close. You mentioned the business case in potentially now retrofitting something to that unit that can make it connected. Does that scale? Do you find that it scales? And are we there yet?

21:52
Speaker 3
Yeah, so let me try and answer that. Starting off with Lenny’s initial question, and then, Jaco, working towards your question of scale. I think it all starts at the point where you need to ask yourself, what is a sensor? Because I think that is where the big paradox actually comes into the equation. If you look at it, purely technical from my, with my engineering mind, you know, the sensor is the device that actually takes the measurement. And if that’s the answer you’re using, then the sensors are actually quite cheap, you know, to the point that Lenny was asking. The reality, however, is that in IoT, the sensor is not just the device that takes the measurement, it is the ability for that device to actually communicate with something else. And that changed the equation entirely. So now let’s look at your example.

22:54
Speaker 3
You were talking about the fridge. Let’s just go back to the fridge example. Jakob, you’ve mentioned now. So it’s not difficult to put a temperature sensor into a fridge, and that sensor doesn’t need to cost more than a couple of hundred granddaddy. But typically those sensors are analogue and they designed to be wired to communicate with. But if you look at the retail store, there’s nothing to wire it to. So you need to find a way of this communicating with a central point, which means this sensor needs to have a way of communicating, typically wirelessly within the environment where it is installed. So that changed the sensor now to a sensor plus a communication module, if you would call it that, comes with it. And that’s where the dilemma starts, because there’s little equipment in the market available.

23:52
Speaker 3
Firstly, that looks like that, the price point suddenly moves up. And thirdly, the moment you bring communication into the equation, the question is, what protocol is used? How does that work? How do you get it to integrate to the other end of the solution that you’re trying to implement? Which raises the question of standardization, to try and get the cost down. And because of this very reason, my own business, in my own business now, we’re actually looking at designing and manufacturing our own, because in the end, that seems to become the only really feasible solution. And that is why you see out there in the market a lot of these point solutions, which is based around the particular sensor set, communicating to a specific platform. Yes, the solution gets designed that way.

24:49
Speaker 3
And once you get scale on the solution, then you can bring the price point down, you know, but if you treat it the way were used to, right in the industrial space, where we basically go in there, we find the problem and we solve the problem, and you try and do the same thing in the more scale environment, such as retail banking, etcetera. You up for hiding to nothing because you have so many endpoints that you’re trying to install. And if you try and solve each one of them individually, yes, it becomes a very costly exercise and it consumes way too much time. You need to find solutions which is simple, quick to install, so that you can use a much less skilled workforce to do the physical installations. And it need to have a fairly standard way of communicating.

25:42
Speaker 3
So you need to make choices. And I think we might get to that a little bit later in our discussion, but you need to make choices with regards to what communication types and protocols do I want to use and how do I limit that? Because there’s so many things out there. We’re talking Zigbee, Sigfox, Lora, you could use Wi Fi, you could use Bluetooth. There’s just a legion of different options that’s out there. Seemingly too many.

26:16
Speaker 1
Seemingly too many options. I think it’s confusing a lot of people.

26:20
Speaker 3
Absolutely, absolutely. And if you do not standardize and get your solution down to one or two of those to cover the use case that you need to deal with, I do not think you will actually get to solve these problems in a way that it will become economically feasible.

26:42
Speaker 1
Quick question and something that I just thought about is we nearly every episode Lenny that’s come up is it versus OT? And I think probably this world and this dynamic has seemingly for the companies that are potentially still lagging and these departments are not working together, the OT and the IT departments is. It typically owns this. However, in the OT world, you know, you have the benefit of coming from that. Both backgrounds having a good experience in both backgrounds. OT typically owns the manufacturing devices and equipment and networks that it would typically own. And there’s always been, always talk about a tongue in cheek. We talk a little bit of a grappling between these two departments. It top of mind is obviously connectivity. Security is a big one for them.

27:38
Speaker 1
On the topic of security, in it and ot, is security perceived as a barrier? Are there still a lot of concerns around security from IT departments? For example, OT teams now want to deploy a whole bunch of connected things. What does that dynamic look like and how does security come into play? And is it still a concern?

28:01
Speaker 3
Yeah, I think it’s a significant barrier, at least a mental barrier, because both environments, both OT and it obviously focus on security, even though their definition of what exactly security might be wouldn’t be exactly the same. But it’s certainly a question asked on both sides, whether you are with Iot or IoT, creating potentially more points of threat in your environment, more points of entry or breach.

28:37
Speaker 1
Correct.

28:38
Speaker 3
And on face value, the answer to that would have to be yes. So, yeah, security would definitely be something that needs to be solved and create some level of comfort with the customer that it’s dealt with. But there’s more than one way of dealing with that, you know, because if we look at security in a physical way, let’s just compare it security with physical security. You know, your typical ADT or somebody that secures your house, what is it that’s really important in terms of security? And in my mind, it is that you need to know that what you’re trying to secure actually has value and actually creates threat to yourself. So if you look at the physical security, you would like to be sure that your family is safe and your assets is safe.

29:44
Speaker 3
We need to think the same way when we think about the IT and OT security. So let’s go back to the example we mentioned earlier, the fridges. It’s getting a bit stale now, but let’s go back to it.

30:00
Speaker 1
A new example. But it’s a practical one, and I like it. I think we can relate easily to that.

30:06
Speaker 3
Yeah, we should actually just think of a new one. But if we look at that one again, it fits the bill, you know, because here’s the real question in that measurement. Let’s, let’s assume that was the only thing that you’re doing in that particular IoT solution. You were only measuring the temperatures of the fridges. It’s a bit simple. It’s often much more complicated than that. But who would be interested in actually knowing the temperature in your fridges apart from that particular retailer who’s actually got value out of that? And what would happen if you do not see it anymore? Now, if you compare that to the OT environment where you were integrating into the physical Plc and the SCADA systems, perhaps that’s controlling the plant.

30:56
Speaker 3
There is a vast difference between the two scenarios, because in the scenario where you actually, on a IoT level, integrating into those systems, the risk or threat is substantially greater in the environment, as I mentioned now, where you’re just measuring the temperature of the fridge, it doesn’t really matter if anybody steals it. So should security at all Beverenne a barrier or a worry? And I think these are the type of questions you need to ask yourself and also use in your business case, as when you put this all together, and by physically designing your systems, in ways to protect where protection is necessary. You don’t need to necessarily worry so generally on every possible aspect of Iot about the security. And I believe in one of your previous podcasts, you had discussions about some of the protocols that used, which included things like MQTT.

32:06
Speaker 3
These protocols already have a fairly significant level of security built in if implemented correctly. So for a very vast number of cases, case studies, if you actually go and look at physical case studies, you would find that security is actually not a big problem, but it is the perception of security which is causing perhaps the barrier, in my opinion. Yeah.

32:37
Speaker 1
Cool. I like that. And I think it does speak to a lot of what we’ve been hearing and some of the conversation we’ve been having, and it has definitely been identified as one of the big barriers. And I think I like that response. And I also like the analogy of comparing it to your own physical perimeter type security that you would have at home and understanding what you need to protect. I want to quickly ask you about, you mentioned platforms. We know that in the OT world with these existing legacy systems that are, you know that typically these organizations have had for several decades the large behemoth systems that are often proprietary in nature and difficult to change, not welcoming of integration. And they’re also very limited, usually in the sense of, in our world we call it tags. Oh, you have ten tags.

33:30
Speaker 1
You can use your ten tags, but anything more than that, you’re going to require a system change or potentially a license change. You can views, for example, you can have ten views, but if it’s more than that, you potentially going to have a system change and a license change. And a lot of those platforms, if I can call it that, or legacy systems, which we now try to create, panel beat into a platform, are very limited in those ways. And you mentioned the barrier of platform. What does that mean? And can you expand on that a little bit?

34:04
Speaker 3
Yeah. So yet again, you get platforms and platforms. If we look at the industrial environment, we’re all very familiar with plant historians, which essentially is a database, time series databases with visualization software and some analysis software running over the top of that. I know I’m very much oversimplifying it, but that is essentially what it does. If you really look at what IoT platform does, it’s pretty much the same thing. However, most of the historians that we are used to was originally definitely designed to run on premise locally at the plant where it’s actually installed and so on. And the models for pricing those, if we can call those platforms as well also evolved out of that environment, which caused them to be things like you’ve mentioned before, a number of tags, a number of views and so on.

35:15
Speaker 3
Whereas in the IoT environment, where you’re looking for scale and quickly upping that numbers, those type of measurements, becomes problematic because it would mean that your cost of the platform would dramatically increase as you’re actually scaling it, which is exactly the opposite of what you want to do. Exactly. Because the idea in IoT is to actually scale it in terms of number of endpoints and keep your cost relatively constant. And by that actually create a margin, a growing margin gap. Yeah. So that you, through scale, keep increasing Margin. So that’s seemingly conflicting. But it’s because we actually comparing the old with the more new today. I do believe there’s some historians and SCaDA systems out there in the market today which has much more attractive billing systems or pricing mechanisms that fits the Iot world better.

36:17
Speaker 3
And this is purely a function of your previous point, Jakku, about OT and it, and how OT and it is actually converging and which then creates more of an understanding, you know, in the OT space, of what’s required. And therefore, a lot of the vendors, the traditional OEM suppliers in the OT space are working very hard on changing their products to be more so we call it cloud friendly.

36:48
Speaker 1
And their models, potentially they have to change the entire model as well.

36:52
Speaker 3
Absolutely. So with doing that, they obviously need to adapt the models and costing models that they use to go with these platforms to actually deal with that. And they’re also entering an environment where they are now competing.