Utilizing Tech - Season 7: AI Data Infrastructure Presented by Solidigm - 05x12: Hardware Beyond the NUC with Julian Chesterfield of Sunlight

Episode Date: July 17, 2023

Although we use Intel's NUC as a shorthand for the type of hardware deployed at the far edge, this recently-cancelled platform isn't all there is. This episode of Utilizing Edge looks beyond the NUC, to platforms from Lenovo, Nvidia, and more, with Julian Chesterfield of Sunlight, Andrew Green, and Stephen Foskett. ARM-based solutions, many using the Nvidia Jetson platform, are particularly interesting given their low cost and power consumption and strong GPUs for edge AI. A hyperconverged stack runs all of the components required for high availability, including storage and networking, in software spanning all of the nodes in a cluster, and this is commonly deployed on low-cost devices at the far edge. The trend to deploying applications at the edge is driven both by new hardware and software capabilities and the changing expectation of consumers and businesses. Hosts: Stephen Foskett: https://www.twitter.com/SFoskett Andrew Green, Analyst at GigaOm: https://www.linkedin.com/in/andrew-green-tech/ Guest: Julian Chesterfield, CTO and Founder, Sunlight.io: https://www.linkedin.com/in/julian-chesterfield-3B74951 Follow Gestalt IT and Utilizing Tech Website: https://www.GestaltIT.com/ Utilizing Tech: https://www.UtilizingTech.com/ Twitter: https://www.twitter.com/GestaltIT Twitter: https://www.twitter.com/UtilizingTech LinkedIn: https://www.linkedin.com/company/Gestalt-IT Tags: #IntelNUC, #EdgeHardware, #EdgeComputing, #UtilizingEdge, @Sunlightio, @UtilizingTech, @SFoskett,

Welcome to Utilizing Tech, the podcast about emerging technology from Gestalt IT. This season of Utilizing Tech focuses on edge computing, which demands a new approach to compute, storage, networking, and more. I'm your host, Stephen Foskett, organizer of Tech Field Day and publisher of Gestalt IT. Joining me today as my co-host is Andrew Green, a veteran of our Edge Field Day and somebody I'm sure you've seen on the internets. Andrew, thanks for joining me. Thanks for having me, Stephen. It was great talking to you last time about Near Edge and Par Edge. Indeed, and that's sort of where we're going here today.

First, though, there's been some breaking news in the industry. As we've talked about for, well, certainly at Edge Field Day and now on Utilizing Edge, the Intel NUC has been one of the sort of poster boy of hardware that's been used for Edge deployment, especially at the far edge. And Intel just announced this week that they are no longer, they're going to be moving away from that business. They're going to be no longer investing in it. I imagine that there is going to be some transition there. But basically, we're at the end of the Intel NUC.

What's your response to that, Andrew? I'm curious what they're going to replace it with, to be honest. Yeah, well, and I don't know if they're going to replace it with anything. They're hoping that their ISV partners will step in and have their own alternative compact form factor PCs. So one of the companies that I've spoken with recently that knows quite a lot about this space and is very aware that the world is not just Intel NUCs. The world is a lot of other PC platforms and alternative low-powered platforms is Sunlight. And we're joined today by Julian Chesterfield from Sunlight. Welcome to the show. Thank you, Stephen.

Thank you, Andrew. Very nice to be here. So I know that, although, as I said, a lot of the industry is really focused on the NUC and that's sort of become shorthand, that's not really as dominant a platform as it seems, right? I mean, there are people using a lot of other devices at the far edge, right? For sure. Yeah. And I think we've seen at the far edge, there's a plethora of different kinds of devices that are more purpose-designed for different solutions. In particular, I would highlight devices like the NVIDIA Jetson platform, which comes in a variety of different form factors

that has embedded GPU acceleration built into it that's ideally designed for computer vision processing and AI workload processing at the edge. I think the edge hardware space is certainly more diverse and interesting, I think, than what we've become used to in the core data center around sort of conventional server platforms and so on. Yeah, there's a bunch of alternative small foreign factor computers. I mean, certainly within the Intel space, there's Lenovo, who have a great range of these things. There's also some, well, there's similar devices from Dell and HPE,

but those are more desktop oriented, at least as far as I know. And of course, then they have their own edge lines that are somewhat different in design from those small form factor PCs. And then, of course, as you mentioned, there's the whole ARM ecosystem that I think we really haven't talked about much. Are you seeing ARM at the edge? Yes, for sure. And, you know, we work with a variety of different hardware platforms, but particularly, you know, Lenovo is a great partner of ours with their, you know, ThinkEdge portfolio. And I think that's quite indicative, actually, you know, Lenovo have a spectrum of devices, you know, everything from, you know from small embedded Intel processors, so the SE10 range, which starts with an embedded Atom processor on it, through to the SE70, which is based on the

NVIDIA Jetson platform. So Jetson, for those who are not familiar with it, is based on it's a system on chip device that has a embedded ARM processor built into the platform alongside the GPU acceleration hardware as well. So, and I think from our perspective and we've worked closely with the ARM architecture for a number of years, I think it's gonna be an interesting space. You know, Arm has a very strong value proposition

for these types of edge workloads, where really the focus is around, you know, AI, data aggregation and fast data processing. You know, a lot of these edge applications that are driving this adoption of technology at the edge are applications that have to be able to do a lot of processing quickly and make fast decisions where latency is an issue. You've got local control operations that need to be executed autonomously at the edge. And so i think um you know this is this is going

to be a strong space for arm um particularly around the uh you know the power consumption so the power footprint um you know we see um that you know often the the the um requirement to to deploy uh power efficient servers and and you know lightweight platforms at the edge is one of the key driving decisions in selecting hardware for applications. So ARM lends itself very well to that architecture, but also the cost points as well. I think when you're deploying infrastructure across thousands of sites in a very distributed manner, cost becomes a really significant factor in doing that. I think one of the factors where organizations are looking to control of those is in the types of devices that they're looking to deploy at the edge.

Those can range from very, very lightweight, low-power, low-data transmission sensors for IoT to fairly intensive computational power for AI and machine learning technology. But how do you exactly scale those resources so you can have a platform that can not only deal with those low-power, low-data transmission transmission ones, the ones that actually do quite a few, quite intensive computation? Yeah, I mean, you know, our approach is, and I think the thing that differentiates us over other solutions that are out there is that, you know, we're a full virtualization platform. So we describe our technology as a true hyper-converged edge stack,

which means that we can run on, you know, very small, lightweight hardware platforms. You know, we can scale down onto, you know, very small sort of mobile device processor platforms, all the way up to larger data center infrastructure as well. I think some of the requirements that we see that drive adoption of sunlight-like technology at the edge is this requirement really, and I think this challenge really grows as your infrastructure at the edge scales out and you have a lot more locations where you're managing infrastructure it's really around you know building a cloud-like platform to be able to manage infrastructure so being able to deploy applications on demand without having to send

people physically to a site in order to install or to manage the infrastructure so you kind of need to have this separation between the infrastructure. So you kind of need to have this separation between the infrastructure management framework that runs on the hardware device and the applications that you deploy on top of that. The environments that we work with don't necessarily have to have a lot of applications running on them.

They might be single purpose devices that run a particular application, but having that ability to manage and control that application out of band from the hardware platform itself is really key. So in the Sunlight stack, for example, we create, because we're a hyper-converged stack, we can provide redundancy and resiliency across multiple physical nodes, which again becomes one of the key requirements for a very large scale distributed edge. That's one of the interesting things I think about a hyperconverged solution at the edge is that, as you say, it's a highly available system. It has a lot of redundancy built in, but the redundancy

is in software, not so much in hardware. And what that means is that you don't have to invest so much in high availability, redundant hardware. You can basically replicate those features in software. And so we've, I think, sort of got this idea that there's sort of this ideal cluster of maybe three low cost nodes running a hyperconverged solution. And that if a node goes out, you can replace them. Is that what you're talking about here? Is the idea that you would have maybe a three node cluster that is redundant and remotely managed? Yeah, exactly. I mean, we can do it with two nodes, which is, you know, kind of an advantage as well. But yeah, it's exactly that. So, you know, we've seen hyperconverged infrastructure in the data center. We've seen, and that's really where the concept

of hyperconverged evolved, you know, folks like Nutanix and VMware that designed the software-defined data center to be able to leverage storage replication, software-defined networking in the data center, all in software rather than having to have dedicated hardware appliances for storing data and replicating data and so on. So those concepts really evolved in the data center. But actually, we see massive applicability at the edge. And I think even more so at the edge, because you have the even greater challenge of in a data center, you often have, you know, smart hands actually on site in the data center where you've got technicians who can go in and help if you've got to pull out disk drives and and, you know, power cycle a device

or diagnose something in the data center, you have that available. You know, you have those sorts of resources available at the true edge um that really doesn't exist um you know you you're talking about um remote um uh you know like in the energy sector remote uh oil fields or uh you know on a bus or in a restaurant or something where um you know it costs a lot of money to actually send people out to these sites. So having the ability to use commodity hardware to some extent, so lower cost infrastructure that you can drop in nodes that look quite similar, but be able to build out that redundancy and that resiliency in the software stack itself is a really key aspect. So what kind of industries and use cases stand out for those edge deployments today? Yeah, that's a good question. I mean, I think there are obviously

some industries that are a bit further down the road with adopting technology and actually bringing in technology and automation into edge environments. The industries that we're focusing on mainly at the moment, or we're seeing the most demand from, I would say are in the hospitality sector, so in sort of retail, you know, supermarkets, quick service restaurants, where, you know, there's this requirement to, you know, margins are really tight, they've really got to try and optimize the cost of running those sorts of businesses. So being able to leverage smart technology, if that's in a restaurant, for example, bringing in kitchen automation, number plate recognition, you know, customer tracking and, you know, and things like that in the restaurant where they can actually bring a,

you know, higher quality of service to people, but at a lower cost ultimately. We're working with use cases in the supermarket sector. There's quite an interesting use case we're developing at the moment with a customer, which involves monitoring refrigeration units in supermarkets where a lot of these devices are quite smart now and the ability to be able to actually monitor and track when devices are not running efficiently or when they're failing and they need an engineer to come out and inspect them. And so being able to bring in technology like that that can make smart decisions across, you know, a large number of IoT devices that are in that location.

The other areas, I think, where the production and extraction of oil is leveraging technology to drive more autonomous decisions and to be smarter about that process of doing that, but also increasingly more in the sort of renewable energy sector. So things like solar power and, you know, wind farms where, you know, in the solar power sector where you've got, you know, lots of solar panels and, you know, the ability to have smart controls to be able to, you know, some of these solar panels can adjust now to sort of follow the sun and so on. And, you know, being able to have localized monitoring and alert detection, you know, triggering alerts if there's problems in those environments.

We're seeing a lot more adoption of technology. And, of course, you know, many of these locations are in quite extreme environments. They need to operate at very extreme temperatures, or they may be out in areas that are not as accessible. So the assumption that everything has high network connectivity, very reliable network connectivity, is often not the case for these types of use cases. So I think having that localized compute logic that can execute on site is very important. And then the other one I would highlight, I think, which is another area where we're seeing a lot of technology adoption for edge use cases is in the industrial sector. So factory automation in particular, being able to enhance quality control,

being able to use things like computer vision apps to monitor safety in industrial locations. So making sure people are wearing the right clothing, they've got hard hats on, making sure they're not walking in the wrong places. If you have delivery trucks coming into locations, making sure that they're unloading safely and loading safely and so on. I think there are a lot of use cases for smart technology

in those environments. And again, that compute logic, I think, needs to be able to run local to where the actual data is being generated in those environments. Do you think that the driver for this is more driven by the new technology trends? You talk about IoT and machine learning and vision processing, solar power, CNC machining, and all these kinds of things. Do you think that that's driving these applications to the edge? Or do you think it's the other way around? Is it consumer expectations and industrial expectations? Basically, we live in a connected world. Everybody's got devices around them all the time. Is that what's driving applications to the edge? it's probably the more so the former because i think the devices are getting more capable

um i think there's um you know there's more um standardization around these devices so it's it's possible to um you know to you know rather than having these closed source systems that have been very much dominated by um you know particular companies that build control systems for these types of sensors and so on. I think what we're seeing now is it's more of an open space where smart solution companies can build software applications that are able to access and store data from these devices and make smart decisions about them. So obviously, there's a demand from the consumer side to have, you know, better safety in these environments or better efficiency to be able to drive more output from an industrial environment. But certainly I think the hardware capability

is probably the thing that's changed most significantly there. It seems like all those use cases that you've described earlier are about industries that have a physical presence like non-digital services for example like your IoT and your like you mentioned refrigeration monitoring computer vision for worker safety it seems like it's really bringing IT and edge into the real world because a lot of other solutions that at least I've looked at for example like CDN based edges, they only talk about digital services.

So how can I make my website faster? Or how can I enable remote working? But it's only a few solutions that can actually deliver those capabilities in the real world. Let's say, how can I deploy this rugged compute hardware into an environment where it can withstand, you know, for example, temperatures of 60 degrees or so?

Yeah, I think, for sure. I mean, it's, you know, it's interesting that, you know, the, I think it is these more sort of traditional Industries almost that are um you know starting to or finding themselves having to modernize in order to stay competitive um and uh you know to to really leverage technology uh to assist with that it's it is you know certainly our experience and, you know, obviously we focus on, you know, I mean, there are particular areas where our technology has most applicability. But I think, you know, we're finding that it is the more the sort of traditional industries that are looking to modernize that are the ones that are actually going to be the earlier adopters of edge technology. I think, you know, obviously there are, you know, other areas, you know, we haven't talked about

smart vehicles and things like that. You know, we have some exposure in those sorts of areas. But I would say that, you know, in general, a lot of the awesome smart automation that's going into, you know, things like self-driving cars and so on is highly regulated. And it's, you know, quite controlled in terms of, you know, the types of technologies that are allowed to be deployed in in vehicles from a safety point of view and so on so it's um uh it it tends to be uh you know and i think we we see that um a lot of the the technology in that industry is is um evolving in a particular direction where you know um many manufacturers are adopting similar approaches to the way they do things. Yeah, that's certainly been our experience.

Well, that's an interesting point you bring up, Julian, because one of the factors at play here as well is non-traditional platforms for edge compute. So I think a lot of us have, again, back to my opening statement about the, you know, the sort of stereotype of, you know, I've got three Intel NUCs in my quick serve restaurant. I think that a lot of this stuff is not being deployed on a platform at all like that. I know that some of this is being deployed in embedded devices, in routers, in industrial settings. We might find hyper-converged infrastructure and multiple hyper-converged nodes being deployed in things that don't look like computers at all. And then you bring up self-driving cars. And of course, there's an entire world

of sort of last mile edge on the consumer side that we've heard some whisperings about as well, deploying things even as far as consumer premises. What do you think of these non-traditional locations for edge platforms? So, yeah, I mean, every environment is somewhat different. I think, you know, if you take the the restaurant use case, you know, the Intel Nook is a good example. It's not a particularly ruggedized platform.

It's something that you can put in a corner, you can stack it up, but it's not something that can operate at very extreme temperatures or it could be, you know, can tolerate a lot of ruggedized movement and so on. It's not really designed for that sort of use case. Obviously, ruggedization comes at a cost. So I think we use this term a lot, the edge-appropriate hardware. And I think edge-appropriate really varies

depending on the type of use case. There is in general, I think a desire in the industry to be able to move towards more general purpose platforms. So being able to deploy onto a more general purpose x86 based architecture or an arm based, you know, if you're a solution provider, having a choice of hardware platforms is smart because it, you know, de-risks to a large extent and also, you know, can reduce the cost. If you've got, you know, if you've got hardware platforms

that are capable of servicing many different types of applications in many different types of applications in many different industries then it means that you know I think everyone can benefit from that sort of efficiency of cost in terms of the the manufacturing process and the you know the the the the cost of the devices themselves I think you know cost is a really significant factor we see in a lot of these edge environments, very much more so than it had been traditionally for enterprises in the data center. I think as we move more out towards the edge and it becomes much more of an operational

part of the business where if you take the restaurant, you know, if you take the restaurant use case, you know, the restaurants operate on very tight margins. And, you know, of course, if you have a large technology cost to actually deploy technology in those restaurants, then that eats into the margin of each of those individual locations. So, you know, being able to find cost-effective solutions, which I think comes down to this desire to be able to deploy on more general-purpose type infrastructure. I want to switch up the gear in a little bit

and ask you about the security aspect. What exactly do we need to be concerned about with security here is it runtime security is it activity or what should we look at exactly yeah good question um i think there are a variety of factors um so you know when you're putting infrastructure out at the edge you know you have to start thinking about physical security for devices. So being sure that a device hasn't been tampered with, being sure that somebody hasn't been able to modify the operating environment for the device, so trusted boot, things like that are important.

But also at runtime itself, you know, having a very tight control over, you know, because now you have a very distributed attack surface from a security point of view, you need to think about, you know, being able to secure the access to the services that are running at that local point, but also being able to trust the integrity of the data that's coming in from your sensor IoT devices or whatever that data source might be. Our approach to the security of the environment itself is, of course, we're a software-based platform. So we work with software solutions that can plug in to our environment. We actually made an announcement not too long ago that we're partnering with a company called Edge Labs,

who have a sort of distributed edge security perimeter type solution, which is, you know, very interesting, very interesting use case for sure and it allows you to across a distributed footprint to be able to to look for different types of attacks that might be and to be able to alert for different types of attacks that might be occurring in any of those sort of edge locations but equally to be able to respond to that and to be able to lock down attacks that, you know, whether it's distributed denial of service or some sort of intrusion it becomes really hard to be able to do that traditional sort of security maintenance in a relatively manual way. You know, you need to rely on the tools and the technology to assist you with that.

So yeah, so I think, you know, security is certainly is a challenge and the more distributed you become, the more you have to think about, you know, how you want to, you know, what are the potential vulnerabilities that you have that you're exposing as part of that? that. One of the design aspects of our technology is that we create a very small footprint. So sunlight is designed to execute with a really tiny footprint, very limited, just a few hundred megabytes of memory, very limited resources itself. But that also extends to the exposure of services. So, you know, we limit the number of services that are actually run and that are actually exposed to the outside. So every edge environment can create a secure tunnel back into the centralized management framework.

And really, it's a trusted relationship between the edge environment and the centralized manager. So we try to really minimize any services that are exposed locally just to the applications themselves, to whatever services they need to expose. Yeah, thank you. You hit the nail on the head there. You have a very wide attack surface area with this. So I think your job as somebody who manages in-edge deployment is to minimize that as much as you can.

And as you mentioned, if you need to have non-technical people installing or working nearby those devices at the edge. You need to make sure that they've not been tampered with, they have secure boot and everything like that. And on the other hand, you need to have your role-based access controls, isolation and segmentation, intrusion prevention, detection systems. So your job here is really, as with any other environment, to minimize the attack surface area and mitigate any vulnerabilities. Yeah, for sure. I mean, role-based access control is another interesting aspect.

In many of the use cases we've talked about, these organizations have different operational teams that might be divided by region. So for example, you might have a North American team that manages infrastructure in a particular area, you might have a European team or you know different different areas globally where you've got folks who are responsible for managing that infrastructure or to responding to issues that might occur in that infrastructure. So being able to have role-based access control across the system that allows you to isolate access to specific areas of resource is another requirement, let's say, for a lot of these types of use cases.

And I would bring up the aspect of this out-of-box experience as well, because as you brought up, you know, you don't have a lot of people there. You need to be able to, as you've said to me personally, previously regarding your solution, you need to be able to take it out of the box, plug in the cable, and it just goes. Exactly, yeah. It's the plug and play. You know, edge has to be really simple. And actually, that's one of our taglines is, you know, edge simplicity. The more distributed you get, the more complexity that can potentially add. So, you know, you need to make it as simple as, you know, we refer to the sort of the home broadband modem model where, you know, we've all become used now to receiving when you sign up, when you subscribe to a broadband service, a device arrives in the mail, you unbox it, you plug in to your phone line or whatever the connection may be, and you apply power to it. Everything else just boots up and it kind of manages itself, or at least you've got some centralized dashboard where you can go in and switch on services and do things like that.

That's really, in our view, that's how the edge needs to be. And, you know, the types of use cases we're talking about where it's got to be as simple as it ships out from the factory, pre-installed and pre-configured, ready to stand up. Now, there are, of course, some exceptions to that. You know, we also work in industry environments where, you know, perhaps for security reasons, it might need to run in an isolated environment or it might need to be, it may be in a location where actually there is no network connectivity. So it has to run completely autonomously. But, you know, a large part of the challenge to delivering edge services,

even in those environments, is the ability to pre-set up the environment and then to be able to drop it in so that it can execute. So really, yeah, that's been a core aspect of our architectural design for solving these challenges at the edge. So I guess to summarize, again, although we often talk about this sort of ideal of, yeah, we're using some mini PCs at our quick serve restaurant, the edge is a very different space than that. And hyper converged, low low-demand solution is a good one, and it's one that we're seeing a lot more of. How would you summarize the state of the edge,

especially the far-edge client world right now, and where do you think it's going? Yeah, I think it's a very exciting space. And, of course, we're hearing a lot of buzz about it in general, in the general sort of analyst space. I think, you know, edge technology is starting to be more widely adopted. And we're seeing a huge plethora of different applications and solutions that are coming along. I think that it's going to take time to evolve. It takes, you know, large enterprises, you know, these are traditional enterprise verticals that, you know, it takes them time to really try out and adopt this sort of technology. And actually one of the things that we see today is some of the initial

requirements for deploying Edge are that you can run both these legacy applications. I mean, let's face it, Edge technology has been around for a long time. It's not something that's suddenly appeared overnight. So there are a lot of legacy applications that run perhaps on some older sort of legacy Windows environments that have been on dedicated hardware running in these locations. One of the requirements that we see is the ability to be able to support both the legacy apps running in a multi-tenant sort of virtualized way, but equally being able to bring in these newer sort of containerized architectures that are, you know, based around, you know, platform solutions like Kubernetes and Docker

and so on that, you know, we can drop onto Sunlight in parallel to the legacy stuff that has to run. And I think that will be, you know, certainly the first adoption of Edge for many of these environments is that sort of getting over that initial hurdle, but then being able to really leverage the power of the Edge and, you know, having a more general purpose, like a distributed cloud type environment

allows them then to innovate and bring in new solutions and start leveraging more technology as and when they can. So there was a big chasm besides running workloads on-premises when you needed it and those newer types of edge technologies. And I think it took us about 20 years to get across them. But right now, I think your solutions and similar solutions are those that actually enable those through edge use cases. And the key here is not only can I run something locally, but can I run something locally and manage it from a central location and have those distributed across wherever my geographical presence is.

And I think that's the key part that's enabling those edge solutions and is bringing technology into the real world. We're not just obstructed like behind a computer. We can actually go out and interact with it. We can be monitored, we can be safer, and we can be faster. Yeah, absolutely. can be monitored we can be safer and we can be faster yeah absolutely i mean the the reality is that um you know and and i'm sure you guys have heard various statistics around this but the reality is that increasingly more and more data enterprise data is going to be generated outside

of the core data center so you know you have this challenge around you know where do you where do you put the compute processing capability to actually manage and process that data and the answer is invariably it's going to be it varies right there'll be a hybrid solution where you know there are use cases where you want to be able to aggregate and process large amounts of data efficiently in a in a core cloud environment um but you know i think having the ability to uh to to um process data on demand at the edge and react to that data is very important so um you know it's it's it's not a straightforward division between, you know, this is what runs at the edge, this is what runs in the core data center. I think it's very much a hybrid environment. And so, you know, Sunlight certainly can help enterprises bridge that chasm, as you say,

between, you know, being able to choose what is the right location actually to run and execute the processing of that data. And our view is that you've got a centralized control component that gives you a single pane of glass, but ultimately you have to be able to manage the applications and the workloads and be able to deploy them in the appropriate place. Yeah, and I would say too, in my opinion, I think that the classic sort of data center concept of here are our servers

and we're running our things on our servers is increasingly going to get blown up by the edge by the proliferation of different types of devices by the requirements of deploying there and i think increasingly we're going to see a a very different infrastructure there and technology like, like you're describing, especially bringing in lower powered alternative platforms, ruggedized platforms, other devices, I think is really going to kind of lead the way toward that. Well, thank you so much for joining us today.

This has been a really fascinating discussion. Julian, I appreciate your thoughts on the changing client footprint at the edge. Where can people connect with you and continue this conversation with you? Yeah, so thank you very much for inviting me. It's been great to join. I mean, if anyone wants to learn more about Sunlight and our solutions, the best place to go to is our website, so sunlight.io.

And we have a lot of information there and information on how to connect with us and to reach out. You can also subscribe to our newsletter, and we can keep you up to date on new technology that's coming and customer use cases as we deploy them in lots of new and interesting locations. And there's also a blog coming up. to take a look at that and read some of the different viewpoints around deployment of Edge and some of the challenges around deploying Edge technology. How about you, Andrew? What are you doing these days? So I'm neck deep in security at the moment, which is always fun. If you want to get in touch with me, you can do it at andrew.green at pressism.co.

Or you can drop me an email at andrew.green at gigaom.com if you have an edge-like solution that you'd like to be featured in one of our reports. Excellent. Thank you so much. And as for me, you'll find me here at Utilizing Tech every Monday on the On-Premise IT podcast, most Tuesdays, and of course, on the Gestalt IT News Rundown on Wednesdays. Thank you for listening to Utilizing Edge, part of the Utilizing Tech podcast series. If you enjoyed this discussion, please do give us a subscription. You'll find us in all your favorite podcast applications, and you'll also find us on YouTube. This podcast was brought to you by gestaltit.com, your home for IT coverage from across the enterprise. For show notes and more episodes, head over to our dedicated website, utilizingtech.com,

or find us on Twitter and Mastodon at Utilizing Tech. Thanks for listening, and we will see you next week.