Join the Insider! Subscribe today to receive our weekly insights
Willem (00:00)
Welcome, you’re listening to the IT/OT Insider Podcast. I’m your host Willem and in this special series on industrial data ops, I’m joined by my co-author David and by Joel Jacob from Splunk, which is now part of Cisco. If you want to hear more about what’s happening in the world of industrial data and AI, don’t forget to subscribe to this podcast and our blog. Welcome Joel.
Joel Jacob (00:22)
Hey, thanks so much for having me. Very excited to be here today.
Willem (00:27)
and welcome David.
David (00:29)
Thank you, Willem. Also happy to be here and happy to be hosting Joel today. He’s the principal product manager at Splunk, working on their OT capabilities. I’m going to talk about that today. Honestly, I first encountered Splunk as a cybersecurity product back in my days when I was still working on industrial cybersecurity. And I must say I’m looking forward to this conversation because I really want to learn
I would say what steps Splunk has taken and why they are stepping also into the OT world on the data side. So let’s kick this conversation off, Joël, with a short introduction of yourself and Splunk.
Joel Jacob (01:14)
Yeah, I think actually me joining Splunk probably reflects more and more investment of Splunk into the OT side. So my background is actually, I went to school for robotics, but when I graduated, there wasn’t really robotics as a big industry as it is starting to become now. So I worked at Lexis doing quality control, manufacturing engineering, launched the RX350 SUV, and that was really cool.
And then I felt, you know, automotive, it wasn’t moving as fast for me. I was always into like technology. So I went to a smart technology company called ecobee. They do smart thermostats, smart sensors, all these things. And that’s where I really learned about. IOT. And I switched also from engineering to product management to say, well, how do you compete with, you know, Google to have nest and all of these other players? So, okay, we’ll partner with Amazon and Apple and things like that. So three years ago, actually.
someone reached out to me from Splunk and actually I didn’t even know anything about Splunk and I started researching and it’s like, yeah, it’s like a security focused company. do data analytics. you know, why, why are they reaching out to me? And it turns out naturally a lot of the customers because of how powerful the underlying platform is, they started using it for OT use cases. And they said, look, we have this cohort, it’s a growing cohort, but we need someone who has that background.
was the IoT space to really enable it to grow. And that began quite a journey. And even like it’s been through a lot of different transitions. And most recently, as you mentioned, Cisco acquired Splunk, which has been a real tailwind as well, because of course they’re a hardware company versus Splunk was a software company. So now we have this total ecosystem of hardware and software in the OT.
David (03:04)
That’s interesting. And of course, Cisco also has this huge footprint in the industry. So I would say to make things a bit more tangible, to make them a bit more understandable for our audience, I’m going to repeat the same question I’ve asked the previous podcast guests as well. So I’d like you to refer back to our industrial capability.
our industrial data platform capability map. actually made that thing, I made the name way too complex even for myself to remember. So let’s talk about capabilities. I think it’s really important because Splunk is entering this domain might indeed not be known by most of our listeners who come more from the OT domain. So if we talk about these seven capabilities.
when we talk about connectivity and the data platform, the broker, the analytics, et cetera, et et Could you walk us through how data moves and just explain a bit like, what does Splunk do at what level and maybe where does it outperforms the competition? Maybe there are other domains where…
you think the offering is less suitable. I’d like to hear that from you.
Joel Jacob (04:34)
Yeah,
yeah. So I think traditionally, as I mentioned, Splunk is a data platform. And I’ll put that on the back and I’ll come back to that. But I think they, you know, data quality, data broker in store, the analytics of that, that’s really like where Splunk’s strength has been. But what I’ve been trying to grow at Splunk is actually the number one category of connectivity. And then of course, analytics at the edge as part of that connectivity. So
I’ll start on that. So one of the challenges I heard from customers is like, how do I get industrial data into Splunk? There’s existing infrastructure, but then you have to integrate a lot of these middleware. I think this is kind of what’s happening in the industry. You have system integrators, you have do-it-yourself enthusiasts, and they’re like, OK, I need to get that data in. So really, what I was thinking is, OK, how do we make that very easy?
And we actually came out with 18 months ago an edge compute platform. So it’s an edge box. It actually has an AI chip. So I’m going to call it an AI box. It’s not a GPU because it is IP rated and thermal energy and efficiency was definitely something we wanted to focus with. So it has an NPU. But what it allows us to do is run more compute at the edge. Applications, we created our own application.
to run on this Edge Hub device as it was called to make it easy to connect to OPC UA, MQTT, all of these standard protocols and alleviate that challenge to get data into Splunk. But then also because we have this computer on the Edge, this rugged device, we can run and partner with people like Litmus who, hey, they offer all of these protocols, they support ARM, so they’re actually becoming a growing partner of us.
ours. And then once I actually started talking to the Litmus folks, I realized they already had a Splunk connector. So, you know, seemingly their customers are saying, hey, we want to send the data into Splunk. So it’s already there. We didn’t even have to do any integration. It’s built into Litmus already. So it kind of like reinforces that like, yeah, we’re strong on the analytics part. We need to grow in connectivity. And one of the ways to do that is definitely partner with leaders in the space.
David (06:33)
Really?
Joel Jacob (06:54)
And then the other area, and sorry, is on analytics. So there’s platform analytics, but then there’s also edge analytics. So because we have this device, we were saying, how can we move some of the analytics from the platform down to the edge? And then you might ask, why would I want to do it on the edge? There’s a couple situations. is, you know, latency criticality. So, hey, it’s like a time sensitive decisions, bandwidth constraints or reliability. It’s in a remote environment. We might not have good connectivity.
to a platform. a couple of these areas really is pushing edge and of course, you know, edge AI is something. But yeah.
David (07:31)
That’s
interesting. actually, this wasn’t a prepared statement, but in any case, the people of Litmus will also be joining this podcast series. So the ones that want to know more about Litmus, should definitely subscribe because they’ll be also one of the next episodes. But I just want to come back to the analytics part because there is…
Joel Jacob (07:41)
How cool.
David (07:56)
the or I’d say providing an engine or providing a device to run code. But there is also obviously the code development itself. What do you do? Do you offer a device which is able to run code or do you also provide these, I don’t know, either just in Python or low code, no code type of environments where you can also, yeah.
Joel Jacob (08:09)
Mm-hmm.
David (08:24)
build those calculations or deploy those AI models.
Joel Jacob (08:29)
Yeah, so this device Edge Hub is kind of like a flagship to show you what you can do. You can use third party equipment as well, but we’re like, how do we make this very streamlined to pair to the platform and kind of just increase time to value, make it faster? And what we do on that device is we offer a container ecosystem. So you can write Python code and integrate it with other container apps, as I mentioned.
But in our Splunk platform, for example, which can run in the cloud or on premise. So we have a lot of air-gapped customers, know, government customers that have to run fully air-gapped. So our platform can be installed on any of the hyperscalers, your own cloud or your own data center. And then in the platform, we have a tool called Machine Learning Toolkit. On top of this platform, you can have many apps. We have like 3,000 apps. So one of the apps is Machine Learning Toolkit.
And we can actually look at all your data and create these predictive models or are some of our customers have used it even for digital twins. And then from, you know, the time that data comes in through an edge device to get to the platform, you can create some inference in the cloud. Obviously you can send that back down to a machine or send it from operators say, Hey, you know, adjust this parameter to make the machine run more efficiently. Wouldn’t it be great to actually take that and push it to the edge? So we built in our platform, the ability to.
OK, operationalize a model, click Export, and then push it down to that container. And then, of course, you might want to add multi-modal inputs. And that’s where, at least today, you push it to the container, and you can still do a little bit more connectivity. But in the future, we want to make a lot more of that low code, no code as much as possible. And I think with generative AI, there’s a, and I’ll talk about that more later, that trend is actually fast being accelerated.
David (10:25)
Yeah, it definitely is. And so this is really, I would say, on the map, the left part.
I assume that you have quite a lot of knowledge on storing data at scale when it comes to our middle capability, the data broker, the contextualization, the store, et cetera. But is security data and sensor data, is that something different or is that the same or are there at least similarities to be found?
Joel Jacob (10:36)
Yeah?
Yeah.
David (11:02)
How did
you move from log data into sensor data?
Joel Jacob (11:08)
Yeah, so I think, you for people who have heard about Splunk, they know that Splunk is traditionally a big security focus company. We’re actually number one in SIEM, security event information management. So that was a focus and also the secondary focus is IT. So observability of applications. Like you said, hey, I want to know why this application went down. Let me search the log data. there’s an error at this time on this day that coincides with the developer.
So Splunk definitely is coming from the IT side. But something that’s very interesting is that, you know, 10 years ago when Splunk was really starting to ramp up as like a company that’s coming to public markets, IT was also quite fragmented, just like OT is today. So it’s very interesting that Splunk had to face the same dilemma of, hey, like there is some standardization, of course, but there’s always these custom flavors and different variants and like, how do we…
How do we stitch this together? And out of that came kind of the Splunk processing language, which is really like, we’re gonna search data on read. So just dump all your data and then this kind of gets into like the three tiers, bronze, silver, gold. Dump all your raw data into Splunk. And then you can search it and then pull out patterns. So regex, which is regular expression to say, hey, you know, this data is coming in for, I don’t
David (12:14)
Okay, yeah.
Mm-hmm.
Joel Jacob (12:36)
temperature sensor. And there’s a couple of sensors and they’re slightly different, even though they’re all MQTT. Well, we can use this intermediate layer of parsing. And then in the search dynamically cross references to different, you know, slightly different sources, it doesn’t have to be in a predefined format. So that was how Splunk really started tying in all of these third party vendors and disparate sources on the IT side.
And especially in security, can imagine like all these firewall logs, you’re not gonna ever be able to, you know, clean all of that into like a silver tier. Some maybe, some is maybe more valuable. And then that’s what we did. And then another learning we had is, as you hear from all the hyperscalers, it’s like, well, it’s expensive now to have all your data in like search. So then we also have what we call like dynamic data. So we have dynamic archive, which basically is like,
David (13:26)
Yeah.
Joel Jacob (13:34)
You can store data in a colder format and then we’ll pull it into a hotter format for search to make sure it’s the most efficient. You could do dynamic data self-storage. You decide on your own. Or you can do dynamic data active search, which is, you know, it’s just more in the hot, hot storage camp. So there’s a lot of innovation from the IT side that I think, you know, as an OT guy, I would consider myself. We really can take advantage of.
kind of leapfrog some of the challenges we’re faced in the OT world. And that’s kind of the beauty of where I see this OT convergence coming.
Willem (14:13)
Joel, just for me also to be clear, is plug positioning itself more like, let’s say an addition to the ecosystem saying like we do have some capabilities and by plugging in on the edge, we can enrich the existing base, let’s say, or are you guys positioning yourselves as let’s say the OT data platform in general for all the needs?
Joel Jacob (14:41)
Yeah, I think definitely, like I’m sure many have other others have said, we’re part of the ecosystem. You know, we’re not going to say, hey, we’re going to do everything. There are other people who just by their focus areas. Yeah, they’ve really like, you know, I mentioned Litmus, they’re building it all these protocols. And that’s not something we’re going to do. So we definitely are part of the ecosystem. And I think where we play a very strong role is on the analytics and search.
David (14:52)
few.
Joel Jacob (15:08)
I’ll give you kind of another example. We recently came out with something called federated search. So in the security side, for example, you may not want to store all your security data in Splunk or in one place. Amazon has a product called Amazon Security Lake. Well, now with federated search, you can actually take Splunk and search data that’s stored in another place. And we kind of developed that partnership. So federation of basically searching data wherever it is.
David (15:31)
Yeah.
Joel Jacob (15:37)
is kind of another tier of analytics. And how do you enable that? Because everyone’s saying, hey, you have to bring it to our platform first. It’s not going to scale as data continues to scale.
David (15:48)
Yeah,
well, first of all, not going to scale. Secondly, it’s going to be very, very, very costly to move and duplicate all that data. But secondly, nobody’s ever going to win that race, right? There’s always another person going to say like, yeah, but our platform is better. You need to bring all our data in our platform. So I think the federated idea in security, that’s one thing.
Joel Jacob (15:54)
Yeah.
Exactly.
David (16:16)
in the sensor data world. I haven’t heard that too often, but I do think it’s actually a really, really good idea because most of the, I would say concepts today, they still somehow have a central data platform where you bring or you pull or push all that data into the data platform. And then that becomes the single source of truth. But yeah, still, even then it’s very, very difficult to build that single source of truth.
do you actually want to? So that’s an interesting one. Maybe to, I would say, somehow try to conclude the capability map discussion. data sharing, data visualization are the features on the right side of the map. Any thoughts on those?
Joel Jacob (17:07)
Yeah.
Yeah, actually, that’s probably another area we excel at. I wouldn’t have thought so, but then I started looking at the visualization of competitors in the industry and it’s like, okay, this is like Windows 90s era. It’s really not modern. So again, because we’re this modern software platform, we developed a product called Dashboard Studio. And it’s similar to like, you have an asset, you can link it to a search. So this asset is dynamically refreshing on the screen.
but it’s coded in a way that even a little things like, you know, dark mode, okay. It’s like such a insignificant thing, but it’s actually like something that’s an easy thing for us to do. Cause a platform is done in a way where it’s using the latest state of the art software, right? It’s not like, that’s a six month project to actually bring that to you. It’s like, okay, we can do that very quickly. And you know, this actually also relates to, where we’re headed also with generative AI. So.
David (17:48)
Yeah.
Joel Jacob (18:11)
I mentioned we use this thing called Splunk processing language. You know, it’s different than SQL, similar, but different where it doesn’t need structured data. We can do unstructured data. Now that everything is a search, and I don’t know if you’ve seen in the biggest trend recently at the time of this recording is code generation for generative AI, right? So cursor, it’s like, hey, you can ask it to create some code. Now, because it’s a coding platform, we can actually,
David (18:32)
huh.
Joel Jacob (18:40)
generate dashboards, you can ask it to generate inferences and results. And not just like create the searches, but also show that to you in a really nice looking way. So it’s going to really shrink the time to be like, Hey, I have this executive presentation. I need to show all the quality defects on this line. You can actually start asking the system. and you know, we’re, seeing a lot of growth. We currently have something called, AI assistant for SPL. So again, you don’t need to learn how to do SPL writing as much.
You just ask the agent and it will start doing the searches for you. So it’s just like, you know, the first step on this early innings we are in the AI, generative AI landscape.
Willem (19:22)
Okay, I think it makes it a bit more clear where Splunk is fitting, let’s say, within the capability maps. We also brought a use case along to make it even more concrete. So maybe we can talk a little about that.
Joel Jacob (19:35)
Yeah, actually, the word you use, concrete, is exactly what I would go with. It’s a concrete, concrete manufacturer. So, you know, what they’ve been doing, it’s actually, again, the security and IT groups, they kind of work with the OT groups and they say, hey, look, we have this really powerful platform. And that’s what’s been able to scale us into the OT side a lot more in the recent years. And the OT engineers are like, okay, we have…
these raw minerals, know, aggregates for concrete and we go and test in the quality lab and you may not believe it, but concrete is actually one of the largest carbon emitting processes in the world, very energy intensive. So for them, any energy savings is super valuable from both the sustainability point of view, but also even like tangible cost, you know, bottom line revenue, because all the electricity they need to power these machines.
David (20:15)
Yeah. Yeah.
Joel Jacob (20:30)
So they put all this data into Splunk from their OT systems, and then they use a tool I mentioned earlier, machine learning toolkit to create predictive models to say, hey, we have a quality lab testing the raw materials here. You know, we have process data on the kiln for the temperature, the speed, the pressure. And then they were able to create predictive models that they’re now able to go down to the PLC and say, hey, change this parameter. We actually don’t need to run it as hot.
And we can still come out with the level of quality we need. Cause as you know, you know, critical infrastructure, you can have concrete that fractures too easily. It’s like bridges. This is a very serious thing. So being able to have confidence that you’ll meet the quality standards, but also be able to optimize your processes. Super valuable for this company that I mentioned is Cementos Argos. And we talked with them last year about the innovations they’ve done.
And they were doing this in the Splunk Cloud platform. And now that we came up with this edge compute device, they’re able to push that model down to the edge. So what it actually allowed them to do typically in the past is they have like, you know, I don’t know, 10 sites or something like that. They were saving $10 million a year with now this energy savings. And now that they can push it down to the edge, it can run 20 % faster. So an extra like 2 million. And then yeah, again, you have like parameters in the PLC to be like, Hey,
these edge cases, if something is really going off, obviously don’t do it. But within these bounds, you know, you can actually automate some of that action. I think someone was mentioning it’s like a closed loop ecosystem, where then operators can get kind of escalated to if there is kind of like, hey, this seems a little anomalous. Can you come check it out? Come come work together. And now it’s this kind of machine and the human working in partnership together. So that’s that’s very exciting for us. And it also opens new
areas of the factory site. So before it was something called the kiln, which is like this really hot process. And then there’s another thing called the control tower where latency is so critical that they couldn’t go there in the past. But now that you have an edge device, you’re able to actually go into these critical environments like, you know, self-driving and these other latency critical things. So yeah, it’s a great reference customer for us. That’s just metric data. But we also, you know, as I mentioned, we’re part of Cisco now.
David (22:37)
Yeah.
Mm-hmm.
Joel Jacob (22:58)
And we have customers
saying, we want to put in camera vision data or sound data. So we have like looking at a blade and listening to the sound of a cutting of paper to determine when the blade is dull or vision. know, there’s like looking at license plates with Cisco Meraki. They have a whole camera suite, but then there’s also like, how do we take that and fuse it with other data source? So we have a camera in a factory and they’re doing torque data and they’re combining it with the vision data. So you’re now like
multimodal, say, right? So like a human in a factory, you have multiple senses. You have your eyes, you have your hands, you can feel if that torque is off. And we’re able to now bring that into kind of the machine. So super exciting. A lot of applications and we’re really just starting to grow, which is probably why a lot of people in the OT place haven’t heard about us yet. But yeah, very much hoping to just make it easier because I’m coming from the industrial side.
And I wish I had a platform that just made it easy, you know, to do these events things.
Willem (24:03)
Is that multimodal part, is that already part of your offering? It’s already integrated or it’s more something on the roadmap to work on.
Joel Jacob (24:11)
It is something we have today. And I think probably the way to explain it is in the Edge device, we have that container ecosystem. So a developer can put whatever they want. So today we actually have people using camera data and metrics data. They have that today, but we want to make that a low code, no code, right? So how do you do that? That’s on the roadmap to make that very easy. But today you can actually, yeah, like we have the data streams, we have the APIs, we have the way to kind of like process it.
So a data scientist, a data analyst can do it today. But a shop floor engineer, you know, hopefully in less than a year, we can release some features and generative AI is actually making us deliver faster. So we can actually bring these features to market faster.
David (24:55)
There is one question I really want to ask. I was going to ask that during the entire conversation, but I see all this cool stuff behind you. I have this very blunt white background today, so nothing fancy.
Willem (25:12)
For the people listening on Spotify or Apple podcasts, go to the YouTube and you can see what’s behind Joel. It’s much more interesting than our backgrounds for sure.
David (25:23)
This is really the moment to turn on your video. What is all the stuff you have behind you, Joel?
Joel Jacob (25:23)
Yeah.
Yeah, so I think as a product manager, a good product manager, you have to really understand the industry. So what I did is this is like my office and I’m like, okay, I need to understand the IT side, because I didn’t even come from the IT side. So I have like full networking in my house, a 10 gig switch, like all of that type of stuff, POE cameras, all that, it’s all network. Yeah, I do have some Cisco gear. I started with some non-Cisco gear and now I’m adapting Cisco gear.
David (25:38)
Yeah.
Is that Cisco gear?
Joel Jacob (26:00)
Because for the personal, it’s too expensive, but for the enterprise, you know? Yeah.
David (26:01)
Yeah, yes. And I don’t
Willem (26:02)
Slightly.
David (26:05)
know whether it’s still the case, but at least in my days, also quite power intensive.
Joel Jacob (26:11)
Yeah,
yeah, exactly. then, and then I have some din rails here. So I’m starting to put like, you know, some ruggedized switches and some other equipment. So hopefully the next time we chat, you’ll see a lot more PLCs and stuff that are in the box that I need to start getting connected. And then, you know, I can just run these experiments at home and really understand like when a user says, this is difficult to use. I can do it myself and be like, okay, yeah, this is, this is a pain point. How do we make this easier?
Willem (26:11)
Electrical heating David electrical heating
David (26:27)
Mm-hmm. Okay.
Joel Jacob (26:40)
So yeah, I don’t want to ever be a person that’s out of touch with the end user.
David (26:46)
I think that’s really cool. And Kate also gave me a couple of ideas. Maybe one day I need to replace my world map, which typically sits behind me with some of that fun stuff.
Joel Jacob (27:01)
Well, you know, we can send you a device so you
Willem (27:01)
Eh.
Joel Jacob (27:04)
can start running some AI models at your home. You’d be surprised.
Willem (27:07)
Usually most people
David (27:07)
Feel free.
Willem (27:08)
also have their networking gear not in their office most of the time. My wife doesn’t like to look at all the cabling and even if it’s well done, she’s not so fond.
David (27:10)
Yeah, yeah.
Joel Jacob (27:12)
Right. Right.
David (27:17)
No, no, no, but actually
Joel Jacob (27:17)
Yeah.
David (27:18)
my, my, and I also have managed to at home, it’s sitting in the basement. So maybe, maybe I could, maybe I could do a picture or so behind me.
Joel Jacob (27:26)
Yeah, yeah, yeah. And well, that’s why I’m here in the basement with the gear. It’s not for the family room.
David (27:34)
yeah, that’s a better idea. We do our podcast recording studio in the basement. we’re now off track. We’re now off track. Yeah, I saw this cool gear and now we’re totally off track. Hey, maybe to get back on track, Joel, just want to quickly go back to your Cement use case. Obviously,
Willem (27:38)
Yeah, I mean, it’s like the IT Crowd version 2.
Joel Jacob (27:42)
Haha.
Yeah, yeah.
David (27:58)
One of the factors here is, or critical factor is collaboration. We talk about IoT collaboration quite a lot. In many companies, you still have this divide between people who are responsible for the PLC side of things, people who are responsible for the data platforms which are running in the cloud somewhere. To do this type of project, you need a form of collaboration. Can you somehow, I would say, comment?
for this case or for other cases like what do you you seen regarding collaboration, what works, what doesn’t, who owns this platform for example.
Joel Jacob (28:37)
Yeah, I think this is a really interesting one. So as a self declared OT guy, I remember this is a little bit of an anecdote. Hopefully we have time. I was working at Toyota and our Toyota Lexus and we had a vehicle where there used to be a black primer that you put on the rear windshield and then it became translucent. So everyone was asking, well, how do know the quality is good? How do know you’re doing it? And we developed a partnership with the IT team.
to say, we’re actually gonna implement a vision system and there’s gonna be a motion sensor, know, PlayStation Move at the time was a thing. So I had a motion sensor on this applicator bottle and actually track in 3D space, putting the primer on the window. And then we could say, all the windows have this good seal and the windows are not gonna fall off while you drive, all this type of stuff. And that was like, my first patent and that was like super exciting. And it kind of got me excited for this OT IT collaboration.
And I found that really it’s successful when you have these kind of innovation teams. And, know, especially now there’s a lot more mandate to be like, let’s get these people work together, put their heads together. At the same time, empathizing with where I was many years ago, I want to also enable the citizen developer, the OT guy. So when we were deciding on this edge device, we actually added like cellular back call as an option. have wifi, ethernet, et cetera. It’s like, why, you know,
There are some remote applications you want cellular, but also I just wanted the OT person to be able to like, okay, I’m just going to set this up, get my feet wet. And then when I know it’s really good, I’m going to bring it to the IT person. And that’s a lot of times what happens. So I’m really thinking about that person who kind of wants to get started and make your platform really accessible to them. But at the same time, when it kind of comes to production and scale, you’re going to have to, you know, you’re going to have to cross that bridge.
Like especially IT is owning security, right? And OT security is huge and Splunk is a security company. So I think one of the selling points is like, you can feel confident that when you do your OT experiments in our ecosystem, you’re not going to run into security issues down the line because we have all the security options and we’re going to check all those boxes. So, you know, really for us, it’s like, how do we excel in OT? And then we can really be at least one platform that helps, you know, make this
David (30:32)
Yeah.
Yeah.
It’s actually a good one.
Joel Jacob (31:01)
divide less far apart, you know?
David (31:02)
Yeah.
I didn’t consider that argument actually because it’s an argument we obviously have faced quite a lot like the security arguments, but you guys just go like, but we’re your security platform anyways.
Joel Jacob (31:18)
Yeah,
exactly. We’ve thought of all of these cases, how to get the security data and like, you know, all of these OT security applications like Stuxnet and all that. That’s like our main business area. So, you know, you shouldn’t be worried about getting into security block with Splunk.
Willem (31:36)
I can also imagine then that your entry points are usually going to be through IT. Like you have like an IT or you’re in an IT organization. How then do you make the step to production? Because I mean, there’s not that many applications that are really used in IT, certainly not in security that make the bridge to the shop floor. So how do you make that link and how do you gain that trust? Because also in the OT world,
Joel Jacob (31:58)
Yeah, totally.
Willem (32:03)
Splunk is not as known as in security, to say the least.
Joel Jacob (32:06)
Yeah, so I mean, there’s a couple things. One, actually, we didn’t mention as much as Cisco. They have a huge footprint. So actually now what we’re doing is, hey, you have a Cisco switch there. It has some limited compute, not as much as the AI box, but we can actually run some applications on your switches. So actually getting into the environment that you’re already connecting your PLCs and everything to, can leverage that Cisco portfolio and make it really easy.
to get data into Splunk. Definitely coming in from the IT side again in that particular use case, but also on the OT side, we actually have system integrators that we’re really working with to go in. And they’re people who know OT. They have OT guys who are doing vibration analytics for robots and predicting downtime and maintenance. They’re either like GSIs, global, the large ones, Accenture, Deloitte, UI, et cetera.
Or they’re like niche ones who are like, I’m focused in oil and gas or I’m focused in, you know, these different verticals. So we’re really working with that group to at least start breaking in more into the OT side. And then, you know, I’m hoping it turns into a little bit of a snowball, as people start seeing like, wow, yeah, this is a really good platform. and one of the interesting things is the next generation of students, I would say they’re more. I T savvy, you know,
So when they go and look at these old systems, they’re like, what is this? I can’t even use this. It’s archaic. And then they come to a modern software system with ours. They’re like, yeah. So we’re hoping that a bit of all of these methods will really, again, close that gap between these two industries that have been wanting to work better together. And hopefully, we can make a way.
David (33:55)
I think that’s a great way to end this episode, Jeroen. The skills gap, the wanting to work together, et cetera. yeah, I would say that’s a wrap for this one, where we again explored how to make industrial data work for us. A big thank you to you, Jeroen Jacob, for sharing your insights and to you, our listeners, for tuning in.
If you enjoyed the conversation, don’t forget to subscribe at itotinsider.com and leave a rating because it really, really helps us. And see you next time for more insights in bridging IT and OT. And until then, take care. Bye-bye.