Edge Computing at the Extreme Edge: Hardware, AI, and High Availability with SNUCs Ed Barkhuysen

Bruce Kornfeld

Welcome to PodMagic, real conversations about solving real IT challenges. I’m your host, Bruce Kornfeld, the Chief Product Officer at StorMagic. And we’re always exploring how simple, reliable technology can benefit you and the people you serve, whether that’s running branch offices, retail stores, or just supporting customers on the front lines.

My goal is to bring super interesting guests, deliver some value and have fun along the way. So, let’s dig in. I’m very excited about today. We have an esteemed colleague, the vice president of product at SNUC, Ed Barkhuysen. Welcome, Ed. Thanks for joining. Can you do a quick intro of yourself and what you do?

Ed Barkhuysen

Yeah, hey Bruce, good to talk to you again and thanks for having me. Happy to be here. So yes, as you mentioned, I am VP of product over at SNUC Systems and we’ve been partnering with StorMagic and many partners. How do we continue to innovate at the edge, bring creative solutions and best support our customers? Happy to be here today. Happy to talk to you and yeah, as you said, we’ve got some fun topics to run through today.

Bruce Kornfeld

I got one quick question that you may or may not want to answer, but can you give us a little bit of background of the change in the company name? Simply NUC, SNUC, was it a pure marketing thing or was there some other rationale behind it?

Ed Barkhuysen

We were known as Simply NUC for many years. So about 10 years old.

And we’re growing and evolving the organization. So, we’re transitioning away from perhaps the traditional NUC business, which was based certainly in the early days in the first years around the Intel NUC platform, which is a fantastic building block to go and generate a number of different possible solutions. So, the NUC product is still alive and kicking. Now it’s being brought to market by ASUS, and we remain a close partner with them. But as the organization grows and as we focus on more Edge, industrial rugged solutions, working with perhaps a slightly different customer base and then growing into more government and federal business. The SNUC systems and the Armadillo logo that we use if you haven’t seen it, represents a little bit of a shift in our company direction and our focus on ruggedness and compute out of the extreme edge.

Bruce Kornfeld

Perfect segue into my next one, which is around ruggedized. That’s what you’re known for. StorMagic is very focused on delivering server and storage and virtualization solutions specifically for the edge. Talk about ruggedization and why are ruggedized solutions becoming more and more important for edge computing environments?

Ed Barkhuysen

I think from a SimplyNUC perspective, we’re known for ruggedized solutions and historically we’ve certainly been known for bringing the maximum level of performance into the smallest possible form practice using our own SimplyNUC technology, but also the latest industry technology. And as we look at those ruggedized solutions and the importance of them, as you move compute out of perhaps a pristine climate-controlled data center into the real world, the rules certainly change and we’re moving with that. And as we look at ruggedized solutions, they are now essential because they are built to survive in conditions that would kill a traditional server. So, think extreme temperatures. And we typically support between zero, minus 40 actually, up to 80 degrees C. Think about heavy vibration on a factory floor or thick dust in a warehouse. So, without a fanless, cableless design, a little bit of debris or a bump could cause a system failure. And that would lead to costly downtime, mission critical spots like oil rigs or autonomous vehicles in particular would not find that acceptable. Now as we look at what we’ve discovered through talking to our customers, many edge compute scenarios are underserved today by the offerings on the market. Often products are taking existing small form factor compute and adding a level of ruggedization around that and not looking holistically at the customer requirement. The typical capabilities of the solutions that are in the traditional data center and how they would be still needed out of the edge. So, from a SNCC systems perspective, we specialize in creating custom solutions for specific industry needs. It could be military or industrial or AI. And we look at often original designs for your compute challenges.

Bruce Kornfeld

Yeah, you know, I was just thinking that’s something that that SNUC and StorMagic have in common is that we see also that on the software side of the world where we play for virtualization, running applications and virtual storage, what we see a lot of times is end users are so used to what they do in the data center. So they’re taking a data center class software and trying to run it at the edge. And that doesn’t work either. So, we focus on simple solutions that are right size for the edge. And it sounds like SNUC is doing the same thing on the hardware side, which is like an interesting parallel that I hadn’t thought of before.

So, you mentioned this. I think we can dig into it a little bit. you mentioned performance needs at the edge. What do you think is different today than it was? I don’t know how long you’ve been in the industry, but let’s just go back 10 years. But what do you think the difference is that’s driving so much more need for compute at these smaller locations?

Ed Barkhuysen

So, I think there’s a couple of real drivers behind this. So, one perhaps might be the, what we perhaps describe as the AI explosion set and AI needs a lot of horsepower. So, you can’t send gigabytes of video footage to the cloud for real time analysis. It’s too slow and expensive. So often we have solutions that are deployed out of the edge with AI accelerators all running natively on the Silicon, the CPU and the hardware as many of the manufacturers are now breaking those NPUs and that capability directly into the products. But also, from an infrastructure simplification perspective, organizations are tired of over-provisioning. They need modular, right-sized compute that does exactly what they need without the bloat of traditional enterprise hardware. And although that’s still absolutely needed in many use cases, now often we’re seeing systems being deployed in harsher, more distributed environments. And customers need hardware that can live there long-term without drama, hence our extreme edge service with that high level of performance in computers there. So, AI is one big driver. But also a lot of the hardware historically out of the edge has been involved in data and transmitting that back, whereas now decisions are being made at the edge, whether that’s perhaps looking at defect control on an industrial line or making real-time decisions perhaps in a retail environment with regards to queue management or perhaps even sort pandemic responses. We saw a lot of that during COVID sort of monitoring at the edge, people and shoppers perhaps complying with the restrictions that the stores needed to have in place. So, lot of that was mandatory and quick decision making out of the edges is certainly a big driver.

Bruce Kornfeld

So you mentioned to me earlier that you might have an example. I don’t know if you call it a case study, but for example technology that you guys have that you’re working on that’s pretty exciting. Do you want to talk about that a little bit?

Ed Barkhuysen

Yeah, that’s for sure. So, I mean, it’s a good example of when we can’t find technology in the market today, which our customers need, and often they haven’t been able to find it either, they come to SNUC Systems. And we’re seeing more and more of this actually as a pattern right now of the organization being a shop which our customers come to when they can’t find anything, a product on the market today that meets their needs. And we can go and create many of these new capabilities and solutions that our customers are looking for.

One good example of that is again, looking at what you might expect to find in a traditional data center. So, as I said, making sure that the solution is right sized and perhaps not just taking traditional compute in a small form back to space and ruggedizing it, but looking at what is traditionally perhaps needed in that server environment and what will still be needed out at the edge. So out of bounds remote management is certainly something our customers are looking for. From SNUC Systems side developed what we call a Nano BNC. So, a Nano Baseboard Controller, which enables full out of bounds control. So, you’ll be able to remotely monitor, manage, patch, update your systems, also introducing full KVM, but this is at the hardware level. So, this is patent pending currently, SNUC systems. And as an example of where, where if something doesn’t exist on the market, perhaps in the form factor in the package that is required to be in, we’ll go and create that and we’ll develop it from scratch and we’ll invest our engineering time and resources to make that happen, to ultimately give our partners and our customers what they need to go and be successful. That’s my example.

Bruce Kornfeld

That’s something where we see the need for that kind of management is that many of our end-user customers are using server technology and obviously our software in hundreds or even thousands of sites and being able to manage remotely down to the server hardware is extremely important. So, it sounds like you guys are innovating in that area as well, which is awesome.

Okay, what about something that we talk about a lot and one of our differentiators is around high availability. So, I think you know this, we’re able to deliver a simple two-node, super highly available solution for customers to run their applications and store their data. What is SNUC Systems view of high availability? Do you see customer demand for it? Talk about what you see about high availability, particularly at the edge.

Ed Barkhuysen

Yeah, so we’re seeing the importance of it growing rapidly. More and more of our business and more and more of our customers are requiring high availability and this is the right language, with mini clusters or micro clusters out at the edge. Previously, as I mentioned, the edge was often just for collecting data. And if it went down, that wasn’t necessarily a disaster. But as we discussed today, the edge is where decisions happen. And if a hospital’s edge node or a smart grid’s controller goes offline, the consequences are immediate and the consequences can be high. So high availability and ensuring there is no single point of failure has moved well beyond the nice to have into becoming a mandatory requirement for many of our edge requirements. So, customers want uptime, but without needing the full enterprise data center footprint to achieve it. So, it’s definitely a big trend that we’re following.

Bruce Kornfeld

Yeah, I’m still blown away. Of course, I’m not going to name names, but we’ve run into some very large end users that have lots and lots of stores out there, let’s say in the retail space. And they’ve been running on a single piece of hardware, a single server. And their rationale in the past has been, if that store goes down, there’s probably another one within five miles or 10 miles. So, the customers will go to the other store.

But we all know that’s probably not true. If I can’t get my McDonald’s hamburger, I might go to Burger King next door. We’re starting to see those companies are rethinking their strategy. Like, we can’t have stores down. We can’t afford that anymore because commerce moves so fast these days. People are making choices. A store being down is a really bad thing. We’re seeing the same thing. High availability is a massive driver for what we do and what our customers are asking us to do.

Okay, so one thing that’s topical that we’re starting to hear more and more about is related to AI, in that all these large data centers are gobbling up all the servers and the GPUs and the hardware. So, we’re starting to see massive spikes on the cost of hardware, particularly around memory. Do you see the same thing? What is SNUC seeing in terms of that? And is that affecting your business? Is that helping your business? Is it not? Give us your thoughts on the recent rise in the cost of memory.

Ed Barkhuysen

From the SNUC systems perspective, we are in a slightly better position perhaps than some others in so much as we have very significant capabilities to bring inventory in-house and that’s helping cushion the impact, not just us, but we can then pass that on to our partners and customers as well. But you are right, this is a huge industry-wide problem. I think there’s various drivers behind it, certainly some of the big CPU manufacturers are focusing on data centers and their higher selling product lines and some of the memory and storage providers are pivoting around to support that as well. And with the end of life of DDR4 and several vendors, the market’s certainly feeling the pinch. And we’re seeing prices going up. So, one thing that we’re doing is being as transparent as we can be with our partners. And as those prices are going up, we’re communicating them. We are seeing it is an industry-wide phenomenon right now.

Really something that I haven’t seen anything quite like this since COVID and obviously different to that as well. And the recovery of it is a lot of speculation about when and if it’s going to recover this year. But from a SNUC Systems perspective, we don’t only have the ability for holding large amounts of inventory and planning ahead, working with our partners to prepare with them and making their plans. But from a purchasing perspective, we have three fully independent but almost copy exact integration facilities as well. So, in the US, in the UK and supporting the EU and the EU as well. So, between those three, we’re able to balance the workload and take advantage of the supply that is available in the market. But yeah, for sure, it is a challenge right now.

Bruce Kornfeld

Okay. Yeah. So as far as what, what StorMagic sees in that area, we’re seeing the same things. Obviously, our answer to the market is what we’re finding customers are doing is they’re slightly delaying some of their purchases for hardware because they can’t wait for the memory to become available or pricing gets a little funky because sometimes we’re talking about thousands of sites. So, one of the things that we do to, to calm customers down is we’re able to run our software on existing hardware. So, if they have one of those more data center, bigger fatter servers available, maybe they leave those in place a few extra months or a few extra quarters but change from a VMware environment to a StorMagic environment. And then they do the hardware replacement a few months later. So that’s something they’re sweating their assets a little bit longer when these things happen. Who knows where this is going? It’s good news for the industry. It’s good news for all the hardware manufacturers because there’s lots of opportunities for everyone to provide solutions for our customers.

Do you have any examples of customer success stories that you want to talk about? Any customer wins that jump to mind that might be an example of where SNUC is one business for extreme edge needs?

Ed Barkhuysen

I think some of the examples I can talk to without perhaps going into all the names and details are really relevant to how do we take compute, putting it as we would say way out at the extreme edge and then making it robust, reliable so it can perform. From an industry’s perspective, there is a lot of demand from defense, so perhaps portable command centers, border enforcement, autonomous transport, and real-time sensor processing that might be more around robotics, navigating a warehouse versus auto motors on the road.

We’re seeing a rise in needs for compute out at the edge where the cloud isn’t actually something which is permitted perhaps in some use cases. And one solution we’re working on fairly recently was around immigration and having translation services enabled through one of our solutions where the translation was completed at the edge, didn’t necessarily want all those conversations being placed into the cloud and processed there. But from an environmental perspective, that’s where we’re really seeing a lot of the focus and drive. So, think airports deploying into secure areas, which are hard to reach, also onto runways. We have products sitting at the end of runways all over the United States and the BMC controller having that extra level of hardware out of bounds remote manageability. Again, that’s really playing into our favor there because certainly you’re going to be closing down a runway if you need to get an engineer on site. So being able to put these systems and reliably robustly remote manage them is critical.

Another interesting one was working on a solution which is deploying right now actually and it’s going into farm yards, into dairy farms and they have a big problem with ammonia and some of their testing processes is to put our products into ammonia chambers and they fill these with 100 % ammonia and they stress test them and they do this for months and they simulate up to seven years plus of being in one of these ammonia rich farm yard environments and they’re happy to say that the products pass flawlessly, not a single failure, so ammonia isn’t a problem for us either.

Bruce Kornfeld

Yeah, I’m just sitting here thinking when you’re telling those stories, I’m just sitting here thinking that like, where the industries come from and where we’re headed, it’s just amazing with the innovation that a company like you guys have, which is ruggedized, small form factor can go anywhere. The innovation for IT people starts to grow. Where else can we capture data? Where else can we analyze data?

It seems like there’s an endless opportunity for you and for all of us, obviously, hardware always needs software too, right? For the whole industry to go find all these new places where data is being created, who would have thought that we’d be measuring ammonia, however you just described it. Whoever thought we’re going to measure these things to that amount of detail.

Think about what’s gone on just in the last five years with Instagram and Snapchat and all the video that people are posting. The innovation is my opinion, I don’t know if it’s fact or not, but the innovation came from Apple, I think, because they were the ones to first introduce high-def cameras into phones. So, everyone on the planet, one billion, two billion, who knows how many billions of phones there are out there, they all are a source of creating content. That didn’t happen just five, seven years ago. Maybe SNUC is the innovator that’s going to help farms and factories and all these other places that hadn’t had servers far at the edge. Maybe that’s going to be all of SNUC’s business. So good for you guys.

Ed Barkhuysen

I think there’s a number of interesting analogies and I’ll just mention, I think from my side as well, like the Apple analogy is good. The iPhone, previously you might have had a camera in your pocket and a wallet and perhaps your keys. And now a lot of that is incorporated into one device into your phone. And I think from your side as well, from a StorMagic perspective, if you look at hyper-converged infrastructure, it’s a similar analogy, right? How do we bring a lot of those traditional capabilities, which would be separated and segregated in data center, but into one reliable platform which we then go and deploy together.

Bruce Kornfeld

I don’t know if the world remembers, we announced this a few months ago, but StorMagic and SNUC announced a tight partnership. Our sales teams, our field teams, our engineering teams, we’re all working together to bring solutions to market. Talk about what the partnership means to SNUC.

Ed Barkhuysen

I think between SNUC systems and StorMagic, we are shaping the next gen of edge compute solutions. And the way I describe it is customers don’t want pieces. They want a complete edge platform that’s easy to deploy and easy to run. So, from SNUC’s side, we bring the rugged, compact, high-performance compute that’s designed for extreme edge environments. And StorMagic brings lightweight virtualization and hyperconverged infrastructure that’s really well suited to distributed sites. So, I think together it solves a big customer challenge. How do I roll out edge infrastructure fast, keep it reliable and manage it remotely without needing a data center team at every location?

Bruce Kornfeld

Absolutely, yeah, we’re really excited about what’s happening here. As I said earlier, hardware needs software, software needs hardware. But between the two companies, think about it, what else does an end user need to run an edge location? Of course, there’s going to be applications on top of us, but the core platform that we jointly provide, small form factor, rugged, can go anywhere, software that’s easy to install, easy to use, simple to manage, super powerful that builds that platform for running applications, storing data is like a really sharp way for us to go to market and solve real customer problems.

Thanks, Ed, for joining, it was a great conversation. I appreciate you coming. Thanks for taking the time. Again, I’m Bruce Kornfeld from StorMagic. You’ve been listening to PodMagic, real conversations about solving real IT challenges.

Ed Barkhuysen

Thank you.