Connecting Offshore AI With Resilient Infrastructure at Tampnet

[00:00:00] Once a year, I open up a handful of places for people who are serious about launching their own podcast, but have never actually gotten around to it. And over the last 11 years, I've helped entrepreneurs, business leaders and Fortune 500 companies all launch their own shows. And here is how it works. You record your conversation on Zoom, Teams or whatever works best for you.

[00:00:23] You send me the audio, an episode title and short description, and I will handle everything else from setting up the podcast host and distribution to getting your show on Apple Podcasts, Spotify and the many, many other podcast platforms out there. I will commit to working alongside you to help you launch your first season of 12 episodes in the bag. So you can start 2027 by ticking a season of podcasts off your bucket list.

[00:00:52] Think of me as your not just personal podcast producer, but also your accountability partner to make sure that you actually deliver those 12 episodes into podcast land. Now, I only offer this once a year with limited places available. So if you want 2027 to be the year you finally launch that podcast, head over to techtalksnetwork.com. Send me a message, send me a DM or even an audio message. I'd love to help you make it happen.

[00:01:20] What happens to industrial AI when the data is being generated on a platform far offshore and then the connection to that shore disappears? Well, the discussion often starts with computing power, but the people operating remote assets, they need dependable communication, predictable response times and systems that behave properly when something fails.

[00:01:46] So on today's episode of infrastructure as a conversation, I'm going to be joined by the CTO of a company called Tampnet. And we're going to discuss the networks connecting offshore platforms, vessels and wind farms. And when sending someone to repair a fault can actually involve weather windows, specialist crews and a journey by sea or helicopter.

[00:02:10] So he will explain today how different communication paths, local processing and traffic priorities all support operations under difficult conditions. And we'll also look at edge computing and AI from video analysis to choosing which connection should carry an application. So if you've ever described an office network outage as a difficult day, I'm hoping today's conversation will give you somewhat of a different perspective on infrastructure resilience.

[00:02:39] Resilience. It's a great one, this. But enough for me. Let me introduce you to him now. So thank you for joining me on the show today. Can you tell everyone listening a little about who you are and what you do? Yeah, my name is Evin Stjavik. I'm the CTO in Tampnet. I've been with the company for the last 11 years.

[00:02:59] Yeah, we are primarily focusing on bridging technology out into the sea, the oil and gas and the maritime environment to provide them with the same capability as you would expect onshore.

[00:03:12] So my job is basically to kind of make some of the world's most industrial remote environment feel digitally connected and secure and responsive, regardless of if it's a fixed platform, an offshore vessel or a wind farm. So it's a big task, but very exciting. It really is a big task. And it's something that we all see on our news.

[00:03:37] We all know that at the bottom of the ocean, these cables there and connectivity, but we don't know too much about it. And one of the things I love exploring on this podcast is demystifying a different area of tech or helping people understand tech that they hear about but don't know too much about. So let's begin with what does reliable connectivity look like when the users and the infrastructure are offshore, remote and incredibly difficult to reach? Good point. I mean, reliability is not just the uptime.

[00:04:07] It's all about having, you know, the consistent performance, predictability of the latency and resiliency of the communication. That's kind of the key component. So our mantra has always been kind of high capacity, low latency. That's our aim, but we are not always able to achieve it depending on locations.

[00:04:30] Sometimes we are relying on satellite, even not only Leo, but also Mayo and even Geo to provide global services.

[00:04:40] So, but again, the main focus is to let our clients feel connected when they are either fixed on an asset platform or a bin farm or even nomadic on a vessel or a drilling rig that commutes across the globe.

[00:05:01] So, so we are, we are kind of relying on a mix of fiber, microwave, satellite and private 4G, 5G system to make that happen.

[00:05:12] And we also focus on multiple transport path to, to, to, to achieve that, obviously, because most of our clients are in business critical and more and more also mission critical state where communication is key. I mean, without communication, they have many, many terms have to shut down basically production, which is obviously not good.

[00:05:42] So it's, it's crucial. Yeah. And just to bring to life some of what we're talking about here, are you able to share maybe a situation where edge connectivity changed an operational decision out there or, or prevented a very costly interruption? You, you, you probably got quite a few. I'm not sure what you can share. You don't have to name any names, but is there a situation you could share there? Yeah.

[00:06:06] I mean, there, there, there's a few during the years that we, we, we can kind of see how communication actually can be looked at. Yeah. As a, as a monetizing element as well to save costs in terms of, uh, uh, not only, not only kind of to, to, to, to fix things, but also equally important to avoid things to happen.

[00:06:27] So, and, and, and, and obviously video analytics, uh, remote workers, and, and all these kind of, uh, connected scenarios, uh, provide, uh, values, uh, and, and, uh, examples across, uh, flare situation, lifting operation, uh, compliance, uh, situation, uh, certificates, uh, has proven to, to, to be solved in, in, in, in, in multiple ways by having, uh, connectivity between, uh, onshore, offshore.

[00:06:57] Having remote hands, uh, tasks to do jobs that before had to be done by on, on-prem people. So, so yes, there, there, there, there, there is a few and, and, and growing day by day as we kind of move up in, in the value chain. Uh, so, um, yeah. And I'm curious when you're, you're in the need there of low latency, high availability and coverage across incredibly harsh physical environments.

[00:07:27] What, what kind of technical trade-offs arise? Anything you can share there? Uh, yeah. Uh, yeah. I mean, uh, physics is always, uh, kind of something we have to, to, to deal with and, uh, coverage and capacity and latency are, are constantly balanced. It's, um, it's, it's, uh, it's an industry of trade-offs because, uh, the frame condition is not always there. So you have to compromise to some extent.

[00:07:55] Uh, but what we typically do is to, to compensate is kind of building redundancy layers, uh, to increase resiliency and, and, and then kind of, uh, achieve, uh, the best outcome. But that obviously also add complexity and cost. So that's where the balance comes in. Uh, so, and, and on top of it, on the physics side, the, you know, environmental, uh, conditions introduce corrosion, vibration.

[00:08:25] Weather, power constraints. So, so it, it all kind of, uh, uh, is, is getting part of the equation. So ultimately the, the, the, the network is engineered for operational outcomes. So it's, it's, it's not just technical specification. It's, it's how you actually, uh, deal, uh, deal with, uh, with, uh, the, the situation when they occur as well. Uh, so a good example of. Yeah.

[00:08:55] Robotics, for instance, they, they rely on, uh, uh, on low latency. Maybe you have a worker communication that, uh, my prioritize coverage and availability and then potentially video analytics, uh, which is more into bandwidth and, and local processing. So, so it depends on the use case, but all in all, it's, uh, uh, uh, built for, uh, or the special purposes.

[00:09:22] So, so when we defined and design a network, we always work close with our clients to understand, first of all, the use cases and, uh, and also the, uh, the kind of, uh, the, the urgency and, uh, and, uh, reliability. They need basically on the SLA to, to, to, to, to, to make these use cases, uh, applicable for the business.

[00:09:53] Yes. Yes. So, so that's, you know, that's what we're doing. Oh, that might insight into a lot of our mindset. network due to active components like for instance the massive MIMO

[00:10:23] they do on 4G 5G. You have to rely on passive component that doesn't give you necessarily the same capacity so if you don't have quality of service underneath the service will suffer. So that's a very key component now when we build sustainable coverage both over the sea and also into the structure.

[00:10:49] And we could have a few people listening in Fortune 500 companies tech teams that feel they've got it tough when they have a P1 incident on a Friday afternoon when it always happens there but I'm curious from in your side of the world here the stakes are much higher how do you design redundancy when repairing a link that could require vessels weather windows and and specialist crews it's a whole other heap of stress levels I would imagine. Yeah yeah obviously you need to

[00:11:18] you need to plan for the unwanted you need to assume failure will the cure. So it's all about designing for continuation rather than recovery. So geographical diversity and multiple transport technologies edge processing to maintain local operation is key especially in the mission critical we are not relying on any connectivity to shore. If you have a mission critical application running there

[00:11:46] it has to be maintained autonomously locally on the asset and obviously the the regular automatic automatic failover and traffic rerouting is kind of in the baseline of the product so so you have to reduce the dependency on a single point of failure basically so so the best repair strategy offshore is often making sure the business doesn't notice default in the first place

[00:12:14] so by having multiple layers of uh of connectivity so to to to kind of achieve the slay that we often deliver as part of our services uh we are relying on on uh to a large degree on on uh on high redundancy I'm curious when you hit i don't know november december january where in norway for example you have

[00:12:39] very limited daylight and a few hours a day there does that make things even more challenging are you able to work around that or can you only work in the daylight? No I mean uh yeah I mean the the all clients are working 24 7 typically so so uh but obviously if you have a blackout then and and you have situation that uh kind of um takes out the regular operation uh environment that might cause a

[00:13:05] challenge uh yeah of course but but we are um again our services are typically it should hit a really hard level if we are completely out uh we are typically degraded uh to the extent that we need to to fix an issue uh in the matter of uh hours days rather than uh seconds and minutes so so obviously

[00:13:32] obviously that that helps us to and and also the remote uh kind of nature of the many of these assets allow us to yeah we are relying on the client to to get helicopter transport or being transported out in the field to to to also do their repairs so so we we are not fully in control of the availability of

[00:13:57] the sites even unless it's a fiber break that where we have full responsibility and and back-to-back agreement with uh with uh yeah repair teams and vessels do you need ai agents that you can trust well with an ai data layer providing real-time connection within your data platforms you can trust your agents to provide accurate solutions so scale your business by trusting your

[00:14:27] agentic ai accurately getting the work done for you trust its capabilities with denodo and you can do that by simply visiting Denodo.com to learn more and i would imagine what makes this even more complicated and occasionally stressful is many industrial organizations are always adding more sensors more automation especially in the ai age now so where do you see networks starting to strain first is there

[00:14:55] anything that you're seeing there yeah typically around the data volumes uh i mean if they increase into dramatically it's it's one thing is the data load itself but i guess the orchestration layer and how to to make sure that you balance out if you have a lot of data points reporting back home at the same time to avoid kind of uh the peak the transmissions to happen you need to kind of divide them into the

[00:15:22] timeline so you have uh you you basically spread the load over time yeah especially on the narrowband and cat m1 and laura van technologies that have very limited uh kind of bandwidth uh even though it collects a lot of endpoints there's not too much uh bandwidth but then it adds up to avoid the the signaling storm and kind of uh all talking at the same time the orchestration is equally important

[00:15:48] as uh uh as the as the load itself uh and obviously video is uh is often the biggest consumer when you start uh talking uh video analytics and everything you can do uh behind that uh that's the big driver for uh uh for bandwidth so especially when you are deploying indoor coverage piggybacking on existing uh distributed

[00:16:17] antenna system and leaky feeders there might be limited capacity so you need again to have a proper quality of service in place to you may make sure you don't constrain the network uh too much and and and and and these days the ot and it convergence is uh maybe one of the the biggest demands because ultimately operation technology is to be exposed towards it to create more situational

[00:16:47] awareness and and have visibility into the operation then you go into a lot of security requirement uh that is very often not uh fully aligned within organization and it's all very often a fight between it and ot organizations to to to safeguard that uh transition the legacy network was were not built for

[00:17:10] ai robotics and thousands of connected devices so so you need to take it in a controlled kind of manner you need to have a good plan behind it so the operational visibility becomes harder when this as the scale increases of course so the bottleneck is often not the bandwidth but but more the ability to manage prioritize and secure

[00:17:37] growing amounts of data i would say uh up against the cyber security element which is uh kind of the de facto most important part when you're exposing uh data between the specialty ot and it and you mentioned exposing data now i would imagine cyber security and physical safety will also meet very closely in these

[00:18:02] environments so how do you you test a connected system will will fail safely and and he could share around that and the the kind of things in place there no again you you have to assume that components

[00:18:16] will fail you have to uh you you have to test degraded operation modes so so what we typically do uh since we are not we don't have the luxury of uh planning with all uh technology and bearers and capacity running at the same time 24 7 so

[00:18:38] so so we define something we call the connectivity modes typically where you have using quality of service to define the level of use cases that you need to support during uh outages so you use all the basically quality of service mechanism to prioritize and even preempt

[00:19:01] mission critical applications that you use all the mission critical applications that you use all the mission critical applications that's the only way you can achieve uh uh mission critical state where you almost can guarantee that uh an important application are getting connected even on uh when you have a when you have a an available kind of um yeah there is in your network so so i would say that's maybe the most important part and and the cyber element

[00:19:31] cyber incident uh cyber incident uh shouldn't create any unsafe operational conditions that that's something you need to plan for you need to have not only cyber security but you also need to have cyber resiliency so in case you have a kind of uh a situation uh compromised situation you need to make sure that you have a kind of uh level of of sustainability in your network

[00:19:57] uh so ot and it segmentation for sure you need to have uh you need to have resiliency across uh all those two technologies and also exercise the failover testing between and uh since we are adhering to the isc 62443 uh the security by design and uh kind of zero trust uh zero trust approach is kind of uh

[00:20:27] uh foundation for for the way we design the solution and then obviously a clear a clear operational procedure when system becomes unavailable how do you act how do you respond uh and how do you distinguish uh within the different uh scenarios and use cases so safety is not only providing the system works it's also

[00:20:52] providing the system behaves correctly when things go wrong uh so to train on that is is key but we we typically don't uh control the application so as a service provider and mno we provide let's say we enable enable the services and we work very closely with our clients to

[00:21:14] to to go through the security posters to set up the connectivity modes as i mentioned and and and make sure that it's in compliance with that is isc 62443 but the way they onboard services and the way they kind of uh operate and manage their uh operation is uh not always within our control but we need to make sure that we have a good foundation for their them to operate on

[00:21:41] and i think very often over the last few years we've heard a lot around ai how it's transforming work for office workers and big companies etc but we don't think about what's happening where you are i mean offshore energy is becoming an increasingly interesting test case for industrial ai i was with ifs recently talking about industrial ai so i know it's a huge um area that's being impacted by ai at the moment but

[00:22:07] how has it impacted you and what's it take to run ai in remote mission critical environments as it dramatically impact your work as well yeah i mean we are we are obviously working in close with ai and we see it as an opportunity as well to extend into new new use cases especially within

[00:22:30] uh robotics drones uh uh video analytics predictive maintenance and and and safety and and uh there's a lot around those new opportunities that we follow closely so so even the edge compute now for us is a new area where we are going into and and extending our services from the connectivity layer also to

[00:22:55] um take responsibility on the edge compute component not only from a workload perspective but also tying it into the into the communication providing you know workloads connected on the device by building policies towards the 4g 5g network we connected to the hyperscalers in cloud by providing secure cloud connection to

[00:23:25] to the hyperscalers we do the van connection for the towards their own data centers and we basically build uh the security posters to um to to run this uh and edge compute is capable of running um not only corporate application but also to to source to host uh repository for uh a lot of the kpis which is selected by video and

[00:23:54] uh and telemetry to to to to be used towards ai inference to uh take decisions and make uh yeah more qualified and faster decisions on their the way they are working and processing uh during their uh operations so it's definitely an opportunity that we follow close and obviously also within the way we operate and uh and build our services ai is important to

[00:24:22] uh to to try to um to simplify things uh to get as close to a closed loop in terms of operational aspect uh detect uh identify and restore services uh in a good way uh and another way of using ai is uh something we call connectivity mix where we have the multi-bearer the multi-home aspect of having

[00:24:49] uh uh an asset connected on uh uh an asset connected on potentially uh line of sight satellite leo meo and 4g 5g we use ai to orchestrate the the traffic uh on a flow-based control between the bearers so so we we identify applications we prioritize applications and we steer them into capacity

[00:25:15] uh depending on their needs so if it's a latency critical application like a voice video it will make sure to to to be routed over a low latency link while a backup and yeah just a file transfer could go on a more high latency capacity link so so that that's uh that's a good way to use ai and also enrich

[00:25:42] uh the decision by uh collecting kpis in terms of signal strength the uh provisioning the bandwidth availability on the on the bears into the decision of where to route the traffic at any given time so so we create a much more dynamic and better flow of uh on the connectivity layers as well

[00:26:07] well we cannot leave without giving a quick shout out to all the people working offshore the unsung heroes that keep things going and it's been great to hear how better connectivity ai robotics how all these things are are bringing remote expertise automation into their daily jobs i would love to give a shout out to all those people and if anyone wants to share their story i invite them to come and join me on here too but for everyone else that would just like to uh find you your team online find out more

[00:26:36] information about tampnet or anything we talked about where would you like me to point them well we do have our tampnet website we are on linkedin for an industry insight and technology updates we tend to join industry conferences and events yeah so any any listeners interesting interested in offshore digital digitalization private 5g edge compute ai and remote operations uh please uh get in touch and

[00:27:05] we will do our best to to work with you awesome well i will have links to everything there and i urge everyone listening to please go check it out it's the largest offshore high capacity communication network in the world there's so much cool stuff over there so i do urge people listening to carry on the conversation with you and also an open invite to everyone working in this field especially around industrial ai and offshore working to to come on here and join me but more than anything thank you for

[00:27:31] raising awareness and bringing this topic to life today really appreciate your time thank you Neil i think having listened to today's guest reliability needs to describe what the operation can keep going when conditions change because a network can lose capacity without every task becoming equally urgent a system can lose its link to the shore while all functions still need to run and my guest explanation today brings those choices into focus

[00:28:01] understand the application agree what it needs test how the service behaves with fewer resources available to it and edge computing also brings a question about where the data should be processed and which decisions need to stay close to the equipment and when you add ai into the mix i think that makes the conversation even bigger while the responsibility for safe operations

[00:28:24] that will always remain real so i'll include links to everything my guest mentioned in the show notes but over to you which services would your organization 100 need to keep running if its preferred connection method became unavailable during operations whether you're offshore or onshore i want to hear from you just tech talks network.com let me know but that's it for today so big thank you as always for

[00:28:51] joining me on the show hopefully you'll join me again real soon for another guest thanks for listening bye for now