Episode 62: The hidden phase of industrial cyberattacks and how to spot it early

Digitalization Tech Talks ·

Show notes

In the 62nd episode of Digitalization Tech Talks, hosts Jonas Norinder and Don Mack kick off a two‑part series on the evolving state of industrial cybersecurity. They uncover what’s really happening inside OT networks long before a cyber incident hits the plant floor including why over 80% of adversary behavior occurs months before impact, why air‑gaps no longer provide real protection, and how legacy vulnerabilities continue to be actively exploited. As guest Adam Robbie puts it

“OT security is no longer a hidden problem — it’s a visibility problem.”

This episode is essential listening for anyone navigating IT/OT convergence, looking for ways to improve industrial defenses, or preparing for emerging cyber risks. The insights come directly from a new white paper supported by a commercial and research partnership between Palo Alto Networks and Siemens, together with Idaho National Labs as the third-party research partner.

Show Notes:

Research Paper (S4 Conference): Intelligence-Driven Active Defense Report 2026 (https://sie.ag/416Tgm)

Website (Palo Alto): OT Security Insights 2025 (https://sie.ag/2ZcgDY)

Website (US Department of Energy): Cybersecurity for the Operational Technology Environment (https://sie.ag/5p9z6a)

Website (Siemens): Industrial cybersecurity solutions | Siemens (https://sie.ag/4HJ8L4)

Contact us:

Adam Robbie email (arobbie@paloaltonetworks.com), LinkedIn (https://www.linkedin.com/in/adamrobbie/)

Tilo Pinkert email (tilo.pinkert@siemens.com), LinkedIn (https://www.linkedin.com/in/tilopinkert/)

Priyanjan Sharma email (priyanjan.sharma@siemens.com), LinkedIn (https://www.linkedin.com/in/priyanjansharma/)

Don Mack email (mack.donald@siemens.com)

Jonas Norinder email (jonas.norinder@siemens.com)

Transcript

So, welcome back to Digitalization Tech Talks. As always, your hosts are Don Mack and myself, Jonas Navinder. And today we're kicking off a special two-part series on the realities of IT and OT convergence and what new research is showing about the current state of industrial cybersecurity. In this first episode, we're going to look at major research efforts that are focused on what's actually happening across industrial environments in a paper that's titled Intelligence Driven Active Defense, Securing Operational Technology Environments. The researchers of this paper analyzed a large amount of real world activity and insights from past incidents, and one theme stood out. A lot happens long before an attack ever reaches the plant floor, and too often organizations miss those early warning signs, even though they can make a huge difference in potentially limiting the impact. We'll talk about what the data reveals about why issues are happening at scale, the early attacker behaviors that show up well before crisis, and how organizations can strengthen their defenses using the framework that came out of this work. Then in our March episode, part two of the cybersecurity series, we'll bring back Marco Ayala to show the practical industry perspective, what OT teams are dealing with day to day, where the blind spots usually are, and how IT and OT convergence is creating new challenges that many organizations are still trying to manage. But before we meet today's guests, and yes, I did pluralize guests, as well as the research behind all this, Don, I'll hand it over to you. Why do you think this conversation matters right now? Jonas, uh, I have a couple of thoughts on this. Uh, first, I think the approach we're taking on this two-part series is a perfect fit for our podcast series. We like to look at the learnings from the recent past to share with our process industries' audience because it might be useful for their own applications. Often we're looking at a specific example with one company, so it'll be really interesting today, I think, to hear about the information gathered across many companies. And specifically on the topic of cybersecurity, I think we both agree that it's becoming an increasingly prevalent issue as we move forward. The deployment of new technology should consider the potential for cyberattacks. So I'm eager to learn from the research that's been conducted by our guests. Yes, me too, Don. So, today we have a full house. Never before have we had 3 guests on the show. So, I have the pleasure of introducing Adam Robbie, the head of threat research at Palo Alto Networks, Brianjan Sharma, senior key expert for security services at Siemens, and Tilo Pinkat, also with Siemens, is senior Director for Business Development and Innovation Management for the Cybersecurity Research Group. So gentlemen, welcome to the show. It's a real pleasure having all of you on the podcast today. Nice being with you. Thank you for inviting us. Thanks for having us. Fantastic. So, why don't we get to it? I want to point the first question or direct the first question to you, Tilo. Could you tell us about what prompted this collaboration between Palo Alto Networks and Siemens and also provide insights on how you combined the latest research paper with Idaho National Laboratory, and also why was now the right time to initiate and release this type of research? That's a really good part because we started the whole collaboration with Palo Alto Networks and Siemens a couple of years ago and actually all goes back already to 2018. And then it deepened and a lot of products and services got integrated and also if you look into the latest more Up to-date partnership events, um, being there on, on, on stages together, so we really work closely to defend and protect uh to defend all the attackers and protect our customers. So it's really working great to have this collaboration and, Already around 2 years ago, I think it was on Adam Robbie who got in more close contact with us and also um for Prianshan, they both started to say like, hey, besides our close commercial collaboration part, why don't we go deeper into our research part because both of the companies, Palo Alto Networks and Siemens, having a dedicated research department. We have it for all the cybersecurity and Palo Alto have it for all the threat research. And so it started actually that we released the first white paper together last January in 2025 and now this year it was time again to come up with a follow up of that one and going even deeper and doing even more. And for doing even more, I hand this over to Adam because it was on Adam to increase the input for the new white paper. With the Idaho National Labs. Thanks, Tilo, and it's really been a very great pleasure working with uh Siemens in the past couple of years and this collaboration is um very important for us as Tilo mentioned that um we are deepening our uh partnership and on the uh threat research level and. Having also a third partner like uh INL and a special thanks to Scott Pond for uh being also one of the main authors helping us to make this happen. Uh, it did help to increase the value and the insight that we are seeing in, um, in the research and bringing more value to our readers as well. Thanks, Tilo and Adam. That was a great introduction. Good to hear about this collaboration. Uh, Pre, perhaps you can take this next question. What were you hoping to learn about the state of industrial cyber risks when you set out to do this work? Um, thanks, John, for the question. Uh, so, with our work, we weren't trying to prove that OT risks exist, because that's kind of already established. We're trying to move past that headline and understand what the risk actually looks like on the ground, and how these cyber risks show up and behaves inside real industrial environments. Now, there's a lot of narrative around industrial cyber threats like nation states, targeted disruption, ransomware. But no large scale empirical data behind those conversations and compared to IT there aren't many well documented cyber incidents we can reference when we talk about industrial environments, and there are several reasons for that, but that's a whole separate discussion. What we focused on instead were a few practical questions, um, like. What kind of threats are actually hitting OT networks? Is it mostly random noise, or are we seeing structured intentional attack behavior? How does the observed activity map to an industrial attack chain? Where does it typically begin and how far it can realistically progress? Uh, another one was, were organizations blind in terms of detection capabilities? And finally, like, what telemetry is truly required to detect these threats, not in theory, but based on what we're actually seeing. And what we learned is that OT environments are definitely not invisible to attackers, and they're far from quiet. There's a significant amount of risky activities happening inside these networks, and more importantly, there's a noticeable gap between monitoring coverage and the actual level of exposure organizations have in place. So, I think it sounds like a very sound approach to study real risks based on actual data. So, based on all this data that you gathered and you analyzed, when you looked at activity happening across these industrial environments, what was it that stood out most about the kinds of risks that organizations are facing today? So here is the breakdown of what stood out for us in these uh analysis, and I would say the top three, main point or one is what we call like the well time. Where We try to know or understand how long it takes from. The initial access for a threat actor or adversary that comes in the network or comes to the environment all the way until it has a huge impact or they completely complete successfully their mission of disruption or ransomware or whatever their purpose are. And as you're going to read in the on the paper, you will see it's it takes like 80% of the adversary activities are happening before their impact, and it takes around 185 days to do that. Then the second part also was very interesting is the The mystery about the air gap, we hear sometimes clients or some companies, they will say like, we don't have that problem or this is threats are not going to impact us because we are air gapped, then. What we did, we looked, we scanned, like, what can we see from outside, what devices, what OT devices are exposed. And we can identify around 20 million OT related devices are exposed to the internet. That includes a systems like an HVAC system and building management system, and those systems, even though they may not sound like our manufacturing devices, but those systems also are crucial in some environment. If HUVAC goes down or the building management system gets locked, right? And also we have seen a huge increment from the last white paper to this white paper on how many devices are exposed, OT devices exposed to the internet. Then the, the last part, which is one of my very interesting points, is the age of the CVEs that's being exploited. We talk a lot about legacy systems. And we know it's a risk. You have a legacy system. It's not bad, you know. We talk also about how many CVEs or vulnerabilities can exist in our environment. That's also an important topic, but what's really high or critical topic we are talking about now is that these CVEs and legacy systems are being exploited, are actively under attack, and we have seen over like 50% of these CVEs that are under attack are aged between 6 to 10 years old. Even worse than that, there's 15% are 16 to 20 years old. So, those brings back to the conversation, legacy systems still exist, air gap is not is a mess, and you have time to protect your environment. A couple of good points, Adam, there, um, relative to A lot of things to watch out for, but good to know that there's a window of time that that could be addressed up front. Um, Adam, if we can continue, I have another question for you. I understand your studies included analysis of over 60,000 firewalls in correlation with, uh, Idaho National Labs research of two decades of industrial incidents. What did that long-term view reveal about how attacks usually unfold and why organizations sometimes struggle to spot the warning signs early? So this is a really great question, like to dive deeper about like what we have seen and why that struggle. I will answer in two points. The first one is the anatomy of the slow motion of the attack, right? What are the TTPs we are seeing? Um, that's been used in these incidents. So what we did, we took a few billions of signatures that's fired on those firewalls, and we mapped them to miter TTPs, and we identified a lots of TTPs are very similar to living on the land techniques. And these shiftings from previous techniques we have seen in the past like remote access or privilege escalation, some techniques are easy to spot, but others are very difficult because they seem legit, but it's still gets hidden in the traffic. Like an example for that, like a command line interface, a CLI that could be used to execute some commands well. It's very difficult to know if this command is legit or malicious if you don't have a really great visibility, and you also identify who is doing what and who's authorized to do what on these devices, especially like in OT environment, it's a little bit easy to do that than IT because you have some expectations, you have some more realistic expectation what to expect from these devices. If you have an engineering workstation on a Windows machine, you Know what to expect, what happened or what behavior to expect from these uh machine. And then you're talking about like the background noise problem, right? So techniques like the native API execution, the standard application layer for command control, those also without like OT specific analytics, these malicious actions look like a routine software update or a standard server query. Those again are very Challenging point to capture and of course the last thing which is similar to what I talked about is the the the concept of air gap. Sometimes operators or asset owners, they may take this for granted, we are air gap, so we don't need to monitor, we don't need to worry so much about what's going on. So, as part of the research, you also outlined an approach that organizations can use to strengthen their defense against some of these things that are maybe not so visible and that might need a little bit more research to figure out what's really going on. Tilo, can you walk us through how that came together and what makes it practical for teams in the process industries? Yes, we looked into the topic and we also looked at how we both are set up for this from our own perspective. For example, how Palo Alto is tackling it for its own topics, how Siemens is going in there for the for its own topics to secure Siemens, and also what we bring to our customers, and we combined it all with this research results we have here. So we gotten to the point that we said, Through the threat analysis, through all the threat research combined with the predictive insights, we generate some nice results, but who needs to handle these results? And there really came our point that we have been diving into the topic around the OTSOC, so a real good secure operational center for OT specific topics. And then we analyzed what we can do about this point specifically. So how could we set it up in a way That it becomes easier to handle that it becomes also more understandable in terms of where are the interactions between IT and OT and where the OTO should really focus on. So we did some analysis on that as well and then we came up with a recommended framework to actually include our results from the research and the predictive insights and come up with an OTSOC framework. What could handle that. And that's the whole idea where we say like if you have all that insights and you have some time to deal with it until you need to react because you know the travel time and you know the average age of the CVEs, basically you also have time to set up an OTSOC and I think that the times on how to walk us through that and um how how how fast it can be to have a certain setup of OTSOCriansion is coming later to that. And for us it was really an idea on how to set it up because there are many papers around OTSOC and there is a lot of input and we thought we really need to focus on some guiding principles on a on a clear vision about it, especially the safety first topic for OTSOC, but we also dive very much into the topic on how to handle governance and operating model around behind it and That is a clear interpretation. So if you're going to have a look into the white paper, you see, for example, where we draw a line for a line for joint accountability. And this line for joint accountability was one thing for us where we really said yes. Below that, it's all about OT stock. You really need to understand your process levels. You really need to understand how your historians are working, what they are producing, what locks they are doing, how that can interchange and actually bring context and what you need to measure and which KPIs you can result from that. And then going up to different levels, you come to the DMC and with the DMC you are coming to this joint accountability issue and we draw. A whole circle around it because we said like if you set it up there's always a next step to develop it a little bit further, do it a little bit better, etc. etc. and it all needs to be guided by the same principles and still needs to be evolving because all the threats are evolving. So this is our way of going forward with that. That is the way we say like this is our recommendation. And to bring all that what we have done here together into the point and saying play active defense with DOToC. Once you have the OT sock there, it's really about you can be proactive from day one because you have time to learn from the, from everything you've got from the predictive analysis, and then you can think about how to set things up and learn from it. And while you're there, you have your time to react. And then you're not just reactive, you really are proactive with the OTSOC framework, and you have all your time to also work on your compliance topics and set this all up. So that's what we combined and that is the way we say we should put forward for all the organizations on strengthening their defenses. So, quite a bit to consider here. And Pre, maybe, uh, Tito had mentioned you were gonna cover this coming up, so maybe we can talk about that now. Um, one of the things we like to talk about for our, our listeners is, you know, what are some of the steps they can take to get started? Could you share the best ways to learn more about your work, explore the framework, and, and begin applying some of these insights inside their own companies? Uh, definitely, Don. So for listeners, they obviously need to start with reading the white paper, but I have like 3 practical steps, which can help them take this white paper in and create some actionable, uh, items from that. So, the first one is, while you're reading the white paper, uh, focus on the model, not just the data. The white paper isn't just about publishing statistics. It's about presenting like a working defensive model. The key takeaway is not that 332% increase in exposed OD devices, but the understanding that over 80% of adversary behavior happened in the precursor phase and long before the operational impact. So the key here really is to understand what this precursor phase looks like for your organization. Second, examine your own exposure versus visibility. The paper emphasizes on threat-informed logging and not the log overload. Use the attach chain models to prioritize telemetry from systems. You don't need every log from each and every device. Understand what matters the most from for your tailored organization. And third, start applying the OTSOC framework incrementally. You don't need a full redesign or transform overnight, but use a phased roadmap approach to the framework. Look at your quick wins, look what you can achieve with the current resources, and then plan the next steps. We do describe a 36 month journey for ODSoC, but it's going to be different for each organization as everyone is at different maturity levels. Now, But what doesn't change is the guiding principles for that Otisoc. And these guiding principles, I guess Tilo hinted towards them earlier, but they start with, what does Otisoc mean? The One thing that OT prioritizes is safety, and the safety first should be the guiding principle of ODSOC. It should be aligned with the business. It should be aligned with the operational, uh, continuance, operational readiness. Uh, it should be complying to regulatory. Frameworks wherever needed and complying to industry standards, which are the organization is following, there should be defense in depth models followed. Uh, there should be effective ITOT integration, which also Tino mentioned is really important to have this joint joint accountability. Now These are the top important things I would ask the readers to keep in mind while they're reading the white paper as they provide like this essential context on how the report should be applied in practice, rather than simply reading it passively. OK, thanks, Pree. That's some good suggestions. And, um, in the pre-release version of the white paper, I had an opportunity to read, you guys did a great job at the end of kind of wrapping up with the conclusions and bringing everything together. And, and in our podcast, we like to do the same thing. We like to end with some key takeaways. …

Full transcript (WebVTT)