This is The Internet Report, where we analyze outages and trends across the Internet through the lens of ThousandEyes Internet and Cloud Intelligence. I’ll be here every other week, sharing the latest outage numbers and highlighting a few interesting outages. This week, we’re taking a break from our usual programming for a conversation about the importance of smooth digital experiences in industrial IoT and robotics. As always, you can read more below or tune in to the podcast for firsthand commentary.
How Networking Orchestrates the World of Industrial IoT
It’s tempting to marvel at breakthroughs in industrial IoT (IIoT) and robotics as if they were isolated milestones in innovation. From agricultural robots tending crops to autonomous delivery robots ushering in new levels of convenience, these days one new capability after another captivates the imagination.
Yet behind every advancement is an intricate web of synchronized communication and constant data flow. If each robot were a musician in an orchestra, each brilliant capability is in of itself a symphony—and modern networking not only delivers the sheet music for this symphony but also conducts it.
In this episode of The Internet Report podcast, Cisco ThousandEyes Principal Solutions Analyst, Mike Hicks, explores the crucial role that communication networks play in enabling innovation in the IIoT and robotics space. We’ll discuss:
- Disruption in the IIoT & Robotics World: Failures or latency in IIoT networks can create high-stakes risks impacting not just functionality, but also physical safety and public trust. The distributed networks that provide critical updates underpin the very social license to operate and further robotics initiatives.
- Data Fidelity, FTW: For IIoT and robotics, network success extends far beyond simple connectivity to making sure that every crucial bit in data transmissions arrive exactly as intended and right on time.
- The Challenges to Monitoring: The distributed nature of IIoT and robotic devices and the impact of environmental conditions on connectivity create a complexity that’s compounded by increasingly agentic behavior. Fixed-destination monitoring is a thing of the past; network operations (NetOps) teams need a holistic view of the complex mesh formed by these interconnected dependencies.
- Where NetOps Must Take IIoT & Robotics: While the future of IIoT and robotics innovation hinges on trust and transparency, the imperative is present right now. Investing in network reliability and assurance is a fundamental business continuity strategy that can directly impact the bottom line and operational integrity.
To learn more, listen now and follow along with the full transcript below.
A Conversation on How Advancements in Networking Power IIoT & Robotics Innovation
BARRY COLLINS: Hi, everyone! Welcome back to The Internet Report where we uncover what's working and what's breaking on the Internet and why. This week, we're talking about all things robotics and the advancements in networking that have powered the innovation in that space. We'll also unpack why smooth digital experiences are critical in the robotics industry.
I'm Barry Collins, and I'll be hosting today with the amazing Mike Hicks, Principal Solutions Analyst at Cisco ThousandEyes. As always, we've included chapters in the episode description below so you can skip ahead to the sections that are most interesting to you.
And if you haven't already, we'd love you to take a moment to give us a follow over at Spotify, Apple Podcasts, or wherever you like to listen.
From robotaxis to surgical robots, we've seen great advances in robotic technology. However, this wouldn't have been possible without advances in networking, would it Mike?
MIKE HICKS: So robotaxis have moved from this novelty to accepted as safe travel. I remember a few years ago, a colleague of ours actually sending me a video of her sitting in the back of this autonomous vehicle as it drove down the road, sort of drove her into work there. Now it's actually really kind of cool, but that was really nothing to do with networking. That was about the basic driving.
Now that happens with the onboard sensors and the local processing. Now, where the networking becomes critical is in that bigger picture: real-time traffic optimization, coordinating with smart traffic lights, fleet management, and route adjustments. So if you think about it, if there's a major accident and the vehicle needs information instantly to reroute, that's when the network becomes critical.
Or software updates. The vehicles are constantly learning and improving, but that knowledge sort of needs to be shared across the entire fleet. The analogy I draw here is, we're thinking about the Formula 1 cars when they're going around, they're sort of having this constant telemetry coming back to make the adjustments, you know, relying on that network. It's not quite the same when we're talking across normal road traffic there, but it has the same implications.
If we take that one stage further, you think about telesurgery. Now these surgeons are performing these procedures from thousands of miles away, so every hand movement absolutely must be transmitted instantly because the surgeon's hands, they're not in the same building as the patient. So these advances, as you say, aren't just driven by robotics alone. Communication networks have played a crucial role to enable this whole ecosystem to interact together rather than just having smart individual machines operating.
So the network really is a difference between a robot that can do its job versus a robot that can do its job optimally as part of this coordinated ecosystem.
BARRY COLLINS: What steps must organizations take to help ensure that their IoT networks are indeed lossless?
MIKE HICKS: So the North Star characteristic is a lossless network, but it's really more about data fidelity and transmission, so ensuring that every bit of that critical information arrives exactly as intended. So it's a two-part equation really. So we have the reliability of the hardware, the routers, the switches, the fiber optics, the cables, the cellular base stations–then plus that, the assurance of bringing that all together. So organizations need to orchestrate communications across a highly dynamic distributed environment. We're not just talking about simple point-to-point connection, but managing multiple data flows, edge computing nodes, cloud services, all simultaneously.
So the key to all of this really is defining what success looks like. So instead of traditional network metrics, robotics SLAs need to focus on that end-to-end service delivery. Can a surgical robot complete a procedure without network-related interruptions? Can the robotaxi fleet maintain optimal routing throughout the whole day? So it's about measuring the business outcome, not just the performance. This all comes back to this fidelity of the data transmission itself. It's not just about monitoring, it's about the need to automatically reroute the traffic, to activate fallback systems, or trigger graceful degradation modes before the robot knows there's a problem, rather than just alerting someone after service has actually failed.
BARRY COLLINS: Tell us more about the impact of potential disruptions or outages in the robotic world.
MIKE HICKS: So network degradations can impact robots differently and it's going to depend on design. All of these systems are going to be built for some sort of autonomous operation and have fallback modes, but network dependent functions will still suffer. So not all these impacts are equal. Connectivity loss has lower impact, say for drone delivery than for surgery, but both affect the trust and overall operational efficiency of this system. So ongoing acceptance and social license to operate these robots is contingent on consistent reliable performance itself. When the robots fail due to a network issue, it doesn't just affect that one task, it damages the public trust in the entire technology.
BARRY COLLINS: Latency must be an important consideration in the world of robotics, especially when you talk about functions such as driverless cars or remote surgery.
MIKE HICKS: You have this concept of latency itself. We talk about this data fidelity of stuff getting there in the time it's intended to. Remember, a lot of these systems are going to have some fallback, so it depends on which function is delayed. If I get information that's sort of delayed getting to me because of a network problem that's critical to redirect a car because there's roadworks in front, it might come too late. I'm actually then already stuck in a roadworks, and I can't then get out of that type of scenario there.
So absolutely, latency has an impact, but it's going to be sort of functionally relevant to a particular task. It'd be remiss of all these people to design a system that, if you lost network connectivity, it just goes completely down and fails. And this is when we talk about this graceful degradation, this graceful slowdown. I have to realize that I'm missing an important factor in terms of the data, and then I need to back off.
But it isn't just about latency. I can have a systems design: all right, I'm buffering the data, I'm doing prefetch of information, I'm constantly polling to understand the traffic signals that are ahead and that type of thing, so I can prefetch that data. So the latency becomes less critical in some senses over the actual fidelity of that data and sort of, the integrity when it comes in.
BARRY COLLINS: Driverless cars are of course reliant on mobile networks, which must bring a whole new set of challenges in terms of assurance, fallback, and latency.
MIKE HICKS: Yeah. It does. And again, you've got to remember that the functions themselves, we're not going to rely on something to do automatic braking. That's going to be within the vehicle itself.
I can't have this problem coming in there, but I do have a network connected into that car for these other systems. And what if I do have then, some critical update that needs to be pushed down to that car? Might just be a road update, might be something with an engine management system or something like that. If there's that sort of delay, you're right.
And now I'm talking across a distributing network, I'm talking about a wireless network, I'm then talking about an on-the-move vehicle potentially that needs to be updated at that time. Now you actually see again, when you design these systems, there's stuff that comes into play where we actually assess the criticality of what needs to be downloaded, and then we'll actually wait until it's stationary so we can have some sort of guarantee over it. So there's sorts of checks and balances in place.
Then if you're sort of going back right to my past in there where you had standby configurations, they also have the ability to sort of roll over to stuff within the vehicle itself. Where it then becomes sort of critical, and if you think about these functions we talk about the surgical stuff where we're driving the arms to perform some surgical procedure there, then that obviously has different criticality. And then that place is where we talk about is the graceful degradation. And obviously, this isn't done completely without any human oversight. At that point then, we'd have sort of people in the room where it's actually taking place, just not the skilled surgeon.
BARRY COLLINS: What challenges are there for monitoring these robotic devices?
MIKE HICKS: There's quite a few challenges.
You have to think, essentially, we're dealing with a distributed architecture and we're effectively dealing with distributed users at this point. Some of these are going to be in fixed places, some of them are going to be moving. If we're talking about vehicles, we might be going to remote locations, and we're certainly having this disparate global one. So the data sources or the input we might need won't necessarily always be local. Now as we said, a lot of it's going to be autonomous, but I have to have that balance between the two, and I have to have this network connectivity.
We then start to consider the environmental conditions. In some situations we're going to have some different environmental conditions. There's the example of the manufacturing robot who's actually sort of checking humidity, therefore to assess whether it's having an impact on the quality of the materials that they're actually using there. I had one situation where an organization was using autonomous forklifts controlled on a wireless network to go and sort of shift the stuff around.
And we had this degradation in performance where we'd actually lose control of the forklifts at different times of the year. And it took a long while to actually work out that what was happening was the Wi-Fi signal was being absorbed by the milk powder that was in the building itself. As the milk powder went down, the signal was able to bounce off the walls effectively and so you could communicate. But when we're actually fully stocked in this warehouse, then the forklift had trouble moving around because the wireless signal was going weak in one particular area where it was going from there.
So these are outside conditions you have to do. So I have to be able to understand what part that plays as well. If we take the autonomous from a robot perspective up into the mine sites or those types of things, again, we can shut down a whole site because we have some sort of condition where the Wi-Fi signal has sort of weakened and gone down. So I need to consider the overarching performance together to sort of bring everything into play.
So it's not just essentially monitoring end to end, it's about monitoring all those nested dependencies and all those dependent systems that come down onto that. And as we move into this agentic mode, where I'm actually going out to seek other information, I don't know where they are. I no longer have a fixed destination where I'm actually getting data from. I might be spilling up a different tool, where I need to actually pull that information from based on a certain set of characteristics that are occurring at that moment in time, be that weather, be that time of day, be that location or you know, have different geographic boundaries. So all of these now have to be taken into consideration. I can no longer look at it in a silo effectively, even for one individual service delivery chain. I have to look at this from a holistic perspective.
I'm having a mesh of these interconnected dependencies. Some of them might actually be sort of disparate dependency, but all of this is going to be relied on to make everything work, and these are all going be connected in some way or form by a network, be that within the wireless network, be that some LTE, be that 5G, then going out obviously into the Internet from a structural perspective.
BARRY COLLINS: What do you think is next on the horizon for robotics and what will NetOps teams need to keep in mind in this new reality?
MIKE HICKS: So the question now is really what can robotics achieve next? And the key to unlocking most of those use cases is going to be really that trust in the performance and that trust in the connectivity. We need to be able to trust that this robot or this IoT device is actually going to be able to perform the action that we're actually asking it to do. But we're also then moving towards this agentic robot.
So systems that sort of dynamically seek out different information sources, then adapt their behavior and even negotiate some of those systems to accomplish these goals. And this starts to add in more complexity, more network data to be able to do that. So think of robots that sort of just don't follow instructions but actively query multiple data sources, to learn from their environment and make an autonomous decision essentially about what information they need.
So as an example of that, if you think of a manufacturing robot on an assembly line going through its processes, through its check, it notices a variation in incoming materials. Instead of just following its programmed responses, it's going to start querying the supplier database to see if there's some sort of quality issue. Then it could check the weather data to see if humidity is impacting materials in its local site. And then cross referencing production schedules, sort of to determine if it needs to adjust its handling technique or alert quality control.
So essentially, at that point it's become a proactive problem solver rather than just a task executor. And what that brings in from NetOps team is we now have again more network interactions. Rather than just sending information down to a system, I have this duplex conversation going up, potentially to other third parties or different dependencies to make this system there. So now teams need to think about these industrial IoT networks differently.
These aren't just data networks anymore, they're mission critical systems to support these intelligent agents. If something goes down on them, it obviously can impact the entire production ultimately the business' bottom line.
The teams need to sort of really design for resilience, redundancy rather than just pursuing zero failure scenarios. It's going to be about this graceful degradation and rapid recovery, understanding when something is starting to fail, be that at a functional level or be that with a component or be that even with some sort of data coming into it, and then being able to understand that's happening and then adjust around it.
BARRY COLLINS: And looking beyond the robotics space, are there any takeaways for organizations and other industries?
MIKE HICKS: So these industrial IoT networks are becoming a significant underlying driver and enabler of advanced automation across all sectors. So in seeing that, organizations need to recognize that their network infrastructure isn't just supporting traditional IT anymore, it's now supporting physical operations. This convergence of digital and physical networks means that these failures are going to have real-world consequences beyond just system downtime. So we're now, potentially, not just impacting productivity, we're now impacting products as well.
And the companies need to evaluate their current network capabilities through this lens–through this lens that they're mission-critical operations, rather than just data transmission. We need to think about this service delivery as a chain, and need to think about this agentic environment, as opposed to just connectivity and availability. The investment in this network reliability and assurance isn't just an IT decision, we’re now sort of moving this into a business continuity decision.
BARRY COLLINS: Any final thoughts you want to leave us with, Mike?
MIKE HICKS: So the future's here. And the technology is now reality, but only if we get those networking foundations right. We're sort of at an inflection point really where robotics technology has proven itself, it's capable of doing it, but it's going to depend on the trust of the invisible infrastructure, which is effectively that networking glue that holds everything together. That's only going to grow, and this is kind of really like sort of the inflection point of when we start to see businesses move and use the Internet as their new WAN.
And this had to build up as a level of trust so they could actually trust that the transmission medium they're actually putting their business on was reliable. We've now seen the same thing effectively within the robotics world or the IoT, the industrial IoT-type environment there.
These robots have to be able to complete their tasks well and without disruption. So we've now moved from this connectivity, “Do I trust my business data onto the network?” to at this point, “Do I have the data fidelity to make sure the robot's going to be able to complete these tasks as we expect them to?”
So it's not about having the smartest robots themselves, it's about having the robots going to work reliably every single time. And to be able to do that, I need to have an understanding essentially of how that network interaction is actually working. It's not just good enough to be able to have connectivity or availability, I need to be able to expand that up into this understanding that the data I'm getting through across this network component is getting there in adequate time and is correct.
BARRY COLLINS: And that's our show. Please give us a follow and leave us a review on your favorite podcast platform. We really appreciate it, and not only does this ensure you're in the know when a new episode is published, it also helps us to shape the show for you. You can follow us on LinkedIn or X @ThousandEyes or send questions and feedback to internetreport@thousandeyes.com. Until next time, goodbye.
