Techstrong Gang – November 26, 2024

Hey folks, I'm Mike Azar. Today we're talking about Mobile Madness. Then we're gonna shift gears and discuss, maybe have we come to the end of Ragtime.

And then finally, well just how green can the Emerald Aisle be in the age of AI and data centers? You're watching Textron Gang, and we'll be back in a minute. All right, folks, it's Tuesday, which means that we're gonna be discussing yet another awesome tech field day.

This one's on mobility. And Steven, I gotta say, you've been doing these things at a pretty hot clip these days. So, um, walk us through the latest and greatest.

Who showed up and what were they talking about? Sure, yeah. So, uh, each tech field day event, we've got different brawn topics.

Uh, as you mentioned, this one's on mobility, which is the, our name for the event that focuses on things like wifi, uh, 5G mobile device management, mobile clients network as a service, that sort of thing. And, um, you know, so each event, although it seems like there's quite a few of them, uh, each one has a different set of companies, a different set of, uh, attendees, a different audience, and so on. Um, this one was, uh, is one of my favorites because mobility and mobile device management, wifi and everything is, it's like a parallel universe.

Um, it's like landing on another planet and, uh, you know, in, in your favorite science fiction book, and you, and you're looking around and you're like, you know, wow, that's, so, that's how they do this, and that's how they do that. And isn't it interesting that they've got, you know, they've invented the same thing in, in a different way. And mobile mobility is like that.

Essentially, uh, everything that we get in the data center, everything we get in the cloud is done in mobile, but it's done with a different flavor because it has to be. So, uh, for example, um, you know, zero trust provisioning and, um, software-defined networking originated in mobility spaces first. Uh, when it became clear that companies would have hundreds, if not thousands of wireless access points, each of which needs to be configured, uh, you know, provisioned, configured, and managed in, you know, by a central device.

And that's, uh, you know, that whole concept has really spread into the cloud. So, uh, this one was no different. Uh, we heard from Cisco, for example.

Um, they're coming in with, uh, a topic that I think is really a, kind of an interesting one. They're talking about wireless backhaul. So for a while in the enterprise wireless space, uh, there, there was a lot of talk about using wifi protocols.

Um, and, uh, that was challenging 'cause wifi wasn't designed for, uh, long distance communication. And so even though there are such devices, they're, they're not, uh, maybe as great as they could be, um, especially when you've got MO devices that are moving around and, and so on. Um, Cisco came in talking about using, um, uh, new wireless protocols to replace, uh, private 5G, which has been another big topic in this space.

Essentially, companies are using exactly the same technology that your phone uses to connect at, uh, high speed to a tower or base station, uh, using that exact same technology to deliver a much, much greater range for wireless devices. Uh, so if you look, think about something like a, um, a farm or a mine or a, uh, big manufacturing facility, they may use private 5G instead of wifi because they need better range. Um, that's also been used for backhaul purposes, and now Cisco's got an alternative solution for that.

Um, uh, a bigger topic, uh, in terms of, uh, what was discussed by the delegates though, was network as a service. So essentially, um, this whole concept, as I said, of, uh, centrally central management and, uh, zero trust provisioning and, and, and modify, you know, monitoring and managing and maintaining networks not as, um, individual devices, but as a enterprise service has really been gaining, gaining traction. Um, we talked about, uh, with a company called Nile that's getting a lot of attention in this space.

Um, uh, a pretty high profile company y'all may have have heard of, and if you haven't, you should probably look it up because I, I think that they're gonna go pretty far. Uh, but the practitioners in the audience were concerned and questioning how exactly this, this is deployed, how it fits in with modern enterprise IT and enterprise networking, and whether network as a service is another one of those things, sort of like, um, infrastructure as a service where it seems like a really great idea until the practitioners actually have to go out there and implement it and manage it. And I think that that's actually gonna come up in some of our later stories here about rag.

Um, you know, maybe we need, uh, to, to, to think about things organizationally, not just technologically. And, uh, so that was another topic. Now another thing the delegates were kind of buzzing about was non wifi wireless protocols.

Uh, many of you have heard about the, the home, uh, standard matter, um, that uses, uh, thread networks, which is, uh, related to ZigBee. There's also Z-Wave and other non-wireless or non wifi protocols. So this is kind of the other end of that, um, wireless backhaul story.

Um, you know, how do you, how do you connect low powered devices always on devices? I mean, here in the office, I use a lot of Z-Wave devices with home assistant, um, as I said, matter and threat are getting a lot of attention. So anyway, if, if, if all of this stuff per piques your interest, uh, do check out some of the mobility field day presentations, but also the round table discussions, podcasts, that sort of thing, because, uh, there's a lot going on in this space.

You know, I've been trying to figure out as of late where that line is between what I, as a member of an IT team or a network admin manage versus that which is managed by the carrier or somebody else. And, um, do you have a vibe for how that's evolving? 'cause you know, back in the day, you know, there were network admins and they were gods and they managed everything.

And then these days it kind of feels like, you know, more and more of this stuff is just a service that kind of is paid for on a monthly basis. Yeah, I think that that's really the, the core of the question here. And like, you know, it really is a, a, a, a question of a trade off between having in-house staff who can manage things versus having things be sort of externally managed as a service for your business.

Uh, some companies, I'm sure would be thrilled to have their networks managed as a service. Um, others, you know, maybe not so much. And I think that just like, uh, when it, when we have discussions about cloud, uh, when we have discussions about platform as a service and software as a service, there's a lot of things that we need to think about that are trade-offs between who's running what and what their SLAs are, and what their responsibilities are in terms of, you know, uh, availability, but also in terms of how to deal with problems when they arise.

And, and I think that some people too, just like in the, the cloud space, are a little nervous about letting, uh, something as essential as their network, uh, be managed by outsiders. Mitch, what's your take on their capability? I'm asking this question because, um, in Boca Raton where Textron is roughly about every 30 days or so, it seems like there's a wireless network issue.

And then, you know, Alan, who's not here, or somebody else is up on a ladder pulling cables and switching things around. And, um, you know, and I scratch my head because, you know, at home here, I don't seem to have a problem with my wifi network. So is there a difference between what happens in the office at home and what's going on here?

The, uh, book office is more of a issue of density, which you run into in different, different, uh, zones within, you know, highly dense, uh, areas in cities and things like that. You don't run into it as much, uh, in, in the communities, in the rural communities. It, I think what's interesting about this, Mike, is that when we say wireless, and now, now it spans multiple things.

It spans what we traditionally thought of cellular wireless, right? When we talk about 5G now, we talk about, uh, cellular, but it's also 5G data. But 5G also includes another, uh, several other types of 5G uh, data networks, uh, that Steam was talking about.

Um, it's been a little while since I've been in the space, but I used to do, had a, had a lab, that lab that did early testing of 5G and software based radios and ran things like that. And also, um, some of the early 5G and then also c certificates for protecting those devices. And, and what's happening is we've kind of blurred the lines now that, that the network, what used to be the edge of the cellular network or the edge of the carrier network, if you want to think of it that way, it's now sitting inside our homes or our offices.

Um, we may get all of our, our service over a cellular 5G type service. We may also use a different type of five, 5G uh, for data networking between buildings or doing long distance, you know, so the, the days of what, where you were talking about, uh, uh, Steven of trying to go long distance with wifi, 'cause that's kind of all we had, you know, as, uh, something less than what a carrier would offer. Uh, enterprises have a lot more options to, uh, pursue.

So your, your, uh, mobile field day is very timely for this. 'cause a lot of organizations try to figure out as they move, move around office locations are changing or closing, and what do they do to collapse or expand the network to the edge? Yeah, and I don't wanna cast aspersions on any particular network administrators, but I will say that the, uh, mobility field, a delegates are some of the experts in this field.

And, um, you know, a, a few of them, for example, are the people responsible for wifi at, uh, major stadiums, uh, factories, that sort of thing. Uh, they know how to make wifi incredibly robust and reliable. Um, back in the day, I remember sitting down with one of those guys and we were having a conversation, and the analogy was, remember when, um, when Windows servers first came out and everybody who had ever used a Windows laptop or desktop thought that they could administer a Windows server because it's Windows, right?

I mean, it's Windows. Of course I can do this. And, um, that was the same that the wifi people felt about non, um, well, non robust enterprise trained people, uh, building and deploying wifi networks.

Um, I've experienced this myself repeatedly, where you go into a place and the wifi network is just horribly misconfigured. Either they've got way too many routers or, uh, or, you know, wireless, uh, access points, way too few. Uh, they're, they're jamming up the spectrum.

They're using too high power, uh, essentially they're, you know, working across purposes. Whereas a, a properly managed and deployed enterprise wifi environment is a, is a thing of beauty. That's, that's all to say though, that, um, all these people have jobs because most wifi networks are not properly managed, properly designed or properly deployed, and basically they suck.

And there are a lot of good people you can hire to do that for you. I know some, Yes, I heard of that. Steven, did anybody talk about AI and the democratization of networking?

Because like I mentioned early on, there was always this notion of the networking gods, but, you know, can mere mortals now master networking and I can just have, you know, your standard IT people manage the network. Well, certainly that's a big topic with a lot of the enterprise, uh, network management companies, uh, both in terms of data center networks, but also campus and mobility networks. Um, lot of talk about using AI to optimize, uh, network designs and so on.

And, and I'm just gonna go out a limb and suggest that probably a lot of that stuff is gonna be moderately successful. Essentially, wifi is, as I said, so poorly managed, so poorly designed that, uh, an artificial intelligence or even, even just a basic template could probably do a much better job than what most people have, uh, at their disposal. So I'm gonna guess that that's gonna be a big wave.

But, um, I guess happily, uh, since it's 2024 and everybody's talking about ai, that wasn't a huge topic at a, at, uh, mobility Field day, simply because, um, it, it's not, it, it's AI is more of a tool for them than it is a, uh, central concern. Mitch, did you ever hear the one about the one thing that the IT admin and the application developer can agree on? It's the network's fault.

It's, it's The networking guy's fault. It's either the network's fault or, or hit the any button, you know, any key, any key on your keyboard, where's the, It's always DNS, It's always dns. You, you know, so, so here's what we get into both, you know, land as well as, you know, wireless type of, uh, networking because of the, the, uh, cable networks, you've, you've heard, seen advertised 10 G by com, by Infinity, uh, Comcast.

That's not a thing. It's not a thing, it's not a thing. It's a marketing label, right?

And, and consciously done that way, I believe. But what's behind that is how they're getting, they're getting much higher bandwidths. It hasn't reached 10 gig, but it, but that's kind of in the scheme of where they're headed.

They do something where they used, um, coherent optics to be able to send multiple, basically optical streams down one piece of fiber. So much like we do kind of spread spectrum or things like that in, in wireless world, they've done that with, with, uh, fiber. And part of that is to be able to be the backhaul for software based radios.

It can be anywhere on the edge. Uh, matter of fact, a lot of the equipment that sits at the edge has been turned into software and could potentially actually live not on the edge, doesn't have to be there. Uh, so it, it's a combination of, you know, they're going that direction.

So what's behind 10 G is not just to your home and to your business, but also as a backbone or back haul for, for, uh, cellular and, and other wireless networks. But I think, you know, I, I learned something long ago. I'm gonna do an Alan, I've been doing this for a long time, and one of the things I've learned is that this is another thing you can always tell who drew the diagram by what's in the center.

If it's the cloud, it was the network guys. Actually, let me, let me follow on that with one last pitch. Uh, if you're a nerd, and you probably are, if you're listening to this, if you're a nerd like me, uh, check out R-T-L-S-D-R.

It's a whole hobbyist community of software defined radio enthusiasts. They're doing just incredible fun things with software defined radios and, um, you know, basic stuff like listening in on the, um, A DSB transmissions from airlines that are flying over to all sorts of just amazing, incredible stuff. It's a whole hobbyist area, and it's a lot of fun.

It's like the ham radio, the 21st century almost. I will say the one thing I keep hearing over and over again is we're building these highly distributed latency sensitive applications, and nobody has any visibility into the networks that they're running over. And so we don't have any understanding of exactly what happens to them when something goes bad.

So, Mitch, do we need to kind of meld maybe observability and networking together a little bit better than we have? I think we do. I think we also need to, you know, what's the saying?

The fu the future is uneven? Well, wireless networks are extremely uneven, uneven, and 5G suffers from all the marketing hype it did about, it's gonna replace everything and gonna be everywhere. And we all know it's not.

And how many times, even if we get a signal, it doesn't, we don't have the data capacity or data connection to Do it. It's not that stable. I have 5G at home and I I, every time I call when there's an outage, it's an external error.

Like, oh, you must live near an airport, or, you know, no one, there's no responsibility being taken. I, I'm glad I have it. It's reasonably priced, but, um, the reliability hopefully will improve.

It's no silver bullet, right? No, but, but it is also, you know, it's coverage and, you know, you're talking about markets outside on the edge of cities and also even dead zones in cities and things like that. So I think that's part of our, our challenge is you in the, in the cable network days, and I did some work in the early, kinda late nineties, mid nineties, and we, we relied on the customers to call us to tell us when the network was down.

That's how we knew there was an issue and customers knew that, right? So they knew to call right away. Maybe there's still some of that going on.

I think we have much better, better visibility, but I don't think we have the observability that we need that you're talking about, Mike, to really understand not just where congestion issues are, but where our coverage issues and where are there, where are there interference issues, right? Well, that's one of the things that people get from buying, uh, you know, and deploying high grade enterprise stuff. So, I mean, if you go to a company like Cisco or HP's, Aruba or Ruckus or, uh, you know, any of these companies, you'll find that their stuff is incredibly robust.

Mm-Hmm. And well instrumented and reliable. Uh, the problem, like I said, it's that a lot of things.

I mean, God, I was at a hotel and they literally had a wifi access point sitting on every desk in every room. And I'm like, well, no wonder your wifi doesn't work. And if you're wondering why your wifi doesn't work, well maybe check out Wireless Field Day Mobility Field day, some of the presentations.

There you go. That's a bad, A more, More is Not better when it comes to wifi. More, not Better wifi.

All right, folks, we're gonna move on the next block, but I'm holding out for six GI think that'll answer everything. Seven G, keep waiting on that, Mike. We'll get to 10 G eventually.

There you go. All right, folks, we'll be back in a minute. Cloud native now is the web's leading resource for the growing cloud native ecosystem.

com is your destination for news, thought leadership, features and webinars on cloud native architecture, Kubernetes, serverless, cloud native application development, microservices, service mesh, cloud native security, and more. Stay on the cutting edge of modern application development at Cloud native. Now.

All right, folks, we're back on the B block here. And we're talking about rag, otherwise known as retrieval, augmented generation for a catchy term for extending large language models into data that they weren't initially trained on. And it requires a significant amount of effort to do that.

You're getting a Vector database or some sort of mastery of, uh, search capability. You gotta set all that up. Typically, you might need a data engineer, and it gets complicated in a hurry.

There's a company out there called Min io that is showing a way to extend the object file system used in the cloud via API to the LLM, so that you don't have to do all that stuff. Steven, I know that you have talked to them at length about this. What's, is this kind of just a one-off kind of solution, or is this the future?

Um, this, well, first off, let me just say yes, I've talked to MIN IO a lot over the years. They've presented at, at Tech Field Day many times since the very fi beginning of the company. And I know the whole team since before min IO was a thing.

So, um, I, I don't wanna sound biased though, because of course we've got a lot of companies that we've talked to. But what that means is that I've talked to them, I've gotten in deep with the product. I've actually used the product in the field.

Um, it is a heck of an object store, essentially Min io. Uh, just to kind of catch people up here, uh, MIN io is probably the world's most popular object store that's not called Amazon S3, and it basically is compatible with Amazon S3. Um, that being said, what we're talking about here is their enterprise product, not their open source product and the enterprise product.

Um, you know, there's, you know, many different ways of, of building a business around an open source product. The way that Min IO has done it is that they've built a very, very robust, full featured, uh, enterprise version of their, uh, S3 compatible object store. And the enterprise version HA is now, um, I believe this is version three.

Um, and it's, yeah, it's AI store. It's basically a, uh, a, a complete revision of the direction of the enterprise products to be an AI based platform and an AI supporting platform. So imagine, I mean, what you know about an object store, what you know about Amazon S3.

Essentially it's a database in the sky, a scalable database. You can put a as much as you want. And we're talking a lot, we're talking many customers with hundreds of petabytes or even exabyte scale data in min io.

And, um, what they've added here in addition to some, uh, you know, decent upgrades that you would see in any next generation enterprise product, so, you know, user interface enhancements and things like that, what they've added is a lot of very cool AI features. Um, and essentially the, my, my, uh, focus, uh, kind of zooms in on two of 'em. Uh, number one, as you described, uh, they have built in a, uh, AI based, uh, query system that allows the, the, this, uh, min io platform to be itself an AI application that can interact with the, with data at that kind of scale.

The other thing that they've built in is, uh, a new protocol that allows these nodes to communicate with each other and with storage or with, uh, client devices incredibly quickly using RDMA protocols. So, I don't know which of those you want to get into. Um, I'm happy to talk about either or both of them.

'cause as you can see, storage nerd, Mitch, you know, I think it was a year ago now, or at least it seems like a year ago, but we hosted a hackathon and we had all these DevOps folks coming in, and you were there for, I think it was like three days. And one of the things that struck me at the time was like how much these developers were discovering the fundamentals of 1 0 1 data management. Like it was a brand new thing.

And, and I kind of like look back at it now, and I'm still scratching my head about it, but, um, are we asking people to know too much about data management and to build these AI applications? And do we just need to simplify this whole thing? Well, since developers are gonna take care of operations and security for us too, we might as well ask 'em to do data management, right?

Yeah. That, well, a lot of, a lot of what we were wrestling with then, then, I guess it's been year and a half ago now, close to is just how do you set up a vector database and how do you structure things so that you can front end, you know, whatever it is you're, whichever model we were experimenting with at the time. So it was kind of a, a science project learning, learning process for us to see how you build these things is really what that, that, uh, that offsite that camp was about.

Absolutely. So, so here's my, here's my view on AI is everybody is, is racing to add AI to their products, which is great. That's fine.

And we're, we're, we've evolving from the chat bot and natural language interface two now AI agents, and there will be more, there'll be orchestrated agents and more kinds of things to come after that. The, the issue isn't adding AI to our, to our product. The issue is, what Min IO is doing is the organizations who effectively leverage their DNA, their data about their company, about their customers, about the market, whatever it is that they, they collect information on to use it, whether it's oil and gas pipelines or it's, you know, at the mall, you know, selling, selling products to people, walking through the stores, the people who effectively find the really interesting and valuable applications for AI to that data and to a business problem.

Those are the winners in ai, not the people just adding AI to their product. Steven, geek out for a minute. How many folks out there have an object file system?

Because last time I checked all these folks were running this unstructured data on something that looked like a file server or a NAS platform or something of that ilk. So how do I get all that stuff into the object file system? Well, uh, you may be surprised 'cause it may already be there.

Um, that being said, you're right that most, I think most companies, if you talk about their unstructured data, well first maybe we start, let's go into reverse here. Um, unstructured versus structured. So structured data is databases, you know, conventional databases, data warehouses, data lakes, et cetera.

And so you've got companies that are doing, you know, some very, very cool stuff there. You know, um, I was just talking to Star Tree about their, uh, incredible, uh, scalable data warehouse. Um, you know, we had Elastic Talk about using, uh, you know, rag integrating, um, vector database access to their, you know, massively scalable databases and so on.

Um, that's all great companies have that, but companies also have what's called unstructured, which is basically a mess, a complete mess. It's all the files and in some cases all the objects that make up everything. And yes, most companies, if you said, what is, where is your data?

That's not your database data, the answer would be, uh, file server. But for the most part, um, I think people also are kind of overlooking the fact that basically every Webscale app uses S3 as their, um, uh, storage engine for blobs of all sorts of different kinds of data S3 protocol and, and, and in some cases actual S3. But of course, there's other companies with compatible, you know, work alike, uh, object stores that work just like S3.

Um, you know, I I I'm familiar with, uh, digital Ocean and CloudFlare and companies like that. Um, so we've got those, those things out there quite a lot. And many companies now have, they're surprised to find that they have more data in an S3 compatible object store than they do on their regular enterprise file store.

Because just naturally what happened was, as you mentioned, the developers, you know, when they're out there developing things, are you gonna, you know, I'm Joe Developer, I'm building a webscale app. Am I going to use NFS or SMB to access a file server to store my data or just dump it in a file system? Or am I gonna call S3?

S3 is much more, um, developer, like, it's much more friendly, it's much more scalable, it's much less administration. So what happens is, in most modern web scale apps, they just, uh, have bucket after bucket of S3 compatible data that's out there. And as I said, many, many of those applications, if they're not using as a service like Amazon's own S3, they're using something that dumps that data into an io um, open source or corporate data store.

So yeah, I think there is a lot more data in, uh, cloud com, you know, S3 object stores than, than you might think. That being said, the other aspect is that in many cases, um, file servers are actually better understood and better managed. And and the same with, for example, Microsoft SharePoint servers.

They're better, better understood than by it because essentially it already has their arms around it. Whereas a lot of that stuff that's getting dumped into these buckets, and it is exabytes of data now that's getting dumped into these buckets, they have no idea what it's, and so for example, Google, um, recently, uh, introduced the ability to basically query your object store and say, what the heck is this stuff? Um, you know, Amazon has some capability like that.

There's a lot of third parties that are doing that. And that's kind of what Min IO is doing here. Essentially, they're, they're giving you the ability not just to say, what the heck is this stuff?

But okay, we've got this stuff now, let me query it, let me ask questions based on this incredible volume of data that's sitting in these giant object stores. I think, I'm Sorry, go ahead. No, I was just gonna say, um, I think overall, um, just bringing this back to the sustainability, um, points, I think that optimizing the storage and having it more centralized is, is, is, is going inherently going to improve sustainability.

I don't know if that was, you know, the objective in this, but I think that it's gonna be something we're gonna see going forward, and we're already seeing it in, in the larger tech companies. So this is another extension of that. It is a big advantage of object stores, you know, stored in buckets as you talked about.

Th those can be anywhere, right? You can place those all over the network. The other thing about this, uh, that I thought was interesting too, Steven, maybe you've talked to them about their AI hub.

I think it's part of the AI sub ai, uh, store story, if you will. But it lets you bring your own models that you might get from hugging face or whatever source, um, that you store in your environment. 'cause oftentimes we don't wanna be going to a public service, right?

We may host those models ourselves. And in effect, it's, it's acting as a rag in front of the object store that has your data in it, as well as interfacing with, with the AI models, gen AI models. So it's kind of its own little ecosystem with a, with a rag and also, uh, storage of your own models that you're using.

So I think that's a pretty fascinating model for enterprises. Yeah, that's actually something that I zoomed into. We talked about this a couple weeks ago on the gestalt, it rundown, and that's one of the things that I zoomed into when they talked about AI Hub to me.

I was like, why are you doing this? Why do, why do, why is a storage an object storage company developing a hugging face compatible, um, model storage? I, oh, there's the word and it, and it's the flip side, right?

So on the one hand of the coin, we have, let me query this, like exabytes of data that I've got in my object store. The other side of the coin is, wait a second, how can I use an object store to better serve ai? And that's actually something, by the way, that Min io presented at our recent AI data infrastructure Field day event too.

There's a, a min IO presentation, you can zoom in there. Um, as I said, they've got S3 protocol over RDMA, which makes it super fast and scalable, but they've also got this AI hub. And, and there again, it all comes down to the fact that a lot of enterprises are using hugging face the same way that they would use GitHub for code.

The problem is that hugging face is a public, you know, software as a service provider. I mean, it, it's essentially, you know, once you put things in hugging face, um, it's, it's a little bit outside your hands. And a lot of people are a little nervous about putting enterprise AI models on something like that.

And so, you know, MIN io did what they did to S3. They basically looked at the, uh, the hugging face, API and they said, what if we built a our own, your own hugging face? So you could store your models, uh, securely locally, um, you know, you can use them internally and uh, to your developers, it looks just like using hugging face, which they like.

And so what it first didn't make any sense to me now makes a lot of sense. Uh, for another reason too, which is that those models take up a lot of a lot of storage space. It's hard to manage a lot of storage, but actually it's not a lot of storage for MIN io and it's actually super easy to manage it on that because it just sort of, there you go.

It, that's, I mean, to MIN io, that's like nothing. You know, they can, they can handle that kind of storage capacity and snapshots and versions, all that kind of stuff, no big deal. Um, so it made some sense to in, to do this, uh, AI hub as well as, uh, some of the other things, All right folks, I'm gonna call it on this subject.

But I would point out that all those folks that grew up in the old school, it ways were taught one thing, nothing good happens when you move data. And I think it's still true. Alright, Hey guys, we'll be back in a minute.

Discover Textron Group, the epicenter of tech innovation. We are your go-to for reaching IT leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us.

You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more. Join our satisfied clients, let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Textron Group.

Hi everyone, welcome back to the Techron Gang. Well, now we're gonna take you on a journey across the Atlantic to talk about a beautiful place. Ireland and Ireland is known for many, many things.

We'll get into that in a little bit. But for one of the things that Ireland has really been, uh, benefiting from and economically is the investment from big tech in terms of data centers and, and lots of planning and, um, things like for expansion of t there. Now, while that has been an economic boom, it's also presented some drawbacks because particularly in the city of Dublin and some others, it is just draining the, the power.

And this is why there is a moratorium actually on new data centers in some of the city limits. So I wanted to take a closer look at Ireland's place I've always wanted to visit, so maybe that's why I was drawn to it. But, um, in the video I'm about to present, I take a look at the data center expansion and the energy demands, and perhaps some potential solutions as we move into 2025 of what Ireland can do to benefit both economically, but also, uh, meet their carbon goals that they have going forward.

Hi everyone. I'm Bonnie Schneider with your Ecotech analyst Insights. Ireland's check Boom has hit a major roadblock, its power grid.

According to a recent report in the Irish Times, there's a growing backlash against data centers due to their energy use. This comes at a time when Ireland is struggling to meet its targets to cut carbon emissions. In Dublin.

Data centers now consume about 14% of the city's total electricity, more than the combined usage of Cork, Galway, and Waterford. Demand shows no sign of slowing down. It's projected by 2030, that number could rise to 30%.

Last year, Ireland's grid operator imposed a moratorium on new data centers in Dublin, effective until 2028. But before assuming Ireland will completely restrict data center growth, it's critical to consider the economic stakes. Data centers have been central to Ireland's tech-driven success, attracting billions in investments from companies like Amazon, Google, and Microsoft.

Looking ahead, Ireland faces the challenge of balancing digital expansion with energy security. This will likely require a multifaceted approach. First, accelerate renewable energy initiatives, particularly offshore wind, which could not only power homes, but also the expanding tech infrastructure.

Second, explore innovations like waste heat recovery. Amazon is already piloting this approach, which repurposes excess heat from data centers to warm nearby buildings. It's a move that could mitigate some environmental impacts.

Finally, instead of clustering facilities in Dublin and risking grid overload, a more distributed model across the country could ease pressure on the power supply and spread economic benefits. The choices Ireland makes in the coming years will have broad implications with global demand for data rising. Other countries are closely watching as they weigh similar trade-offs between economic growth and environmental sustainability.

So digging into renewable energy as a power source is one solution, as I mentioned, as well as looking at other areas to potentially develop these data centers. But it is kind of a tug of war going on right now in Ireland. So that's why I wanted to take a closer look at it.

And of course, we have our Irish expert here, Mike Bazaar to comment. All right, well, both my parents come from Ireland and I still hold citizenship in here and there, but it is a hot debate because the issue is, I think they're on a projection that said, if they actually built everything they approved for data centers, it would consume about 70% of all the power that Ireland is able to generate today. And that would increase pricing.

And there's a huge debate right now in Ireland over housing and the cost of electricity. And there's an election that's coming up on this whole issue, and it's not the kind of topic that's very popular. And some people would point out that there's already more than 80 of these data centers and they've only added maybe 16,000 jobs.

So people are kind of scratching their head going, you know, what exactly is the benefit? I mean, I think the number is, they said that injected 7 billion into the local Irish economy, but outside Ireland, this is the same debate that's happening everywhere, right? Yeah.

We're seeing communities all around the world having the same conversation. These data centers are also a little bit on the noisy side. They don't exactly insert themselves well into the landscape, and we're not quite clear how energy efficient AI will eventually be.

But Steven, I know we've talked about this topic in the past, it's becoming, uh, not just a technical issue, it's a political issue. Yeah. And, and I think that, uh, Bonnie's spot on in, in her analysis here in that, um, yes, data centers take a lot of power, but uh, also the fact that they're so concentrated in the Dublin area is, is a real problem.

And I think that that's what we're seeing globally where, uh, data centers were initially built. I mean, Northern Virginia, for example, the San Francisco Bay Area, um, they are extremely, extremely concentrated in a, in a few spots, but over time they've been, uh, distributed. And of course, Ireland has a lot going for it in terms of data centers.

'cause number, it, it, number one, it, um, it still part of the European Union. Um, it is located, uh, as close as pretty much any other part of the European Union is to the United States. So a lot of transatlantic cables land in Ireland.

Uh, there's a lot of cables as well that go from Ireland to the uk, Ireland to France and so on. Um, but if you look at the map, which I actually did in preparation for this story, those cables don't all land in Dublin. Yes, a lot of them do, but a lot of them are in other places.

So for example, there's a cable landing in Kalala Ireland from New Jersey. Uh, there's a, a cable in, uh, Galway. There's, uh, cables that go to the UK that are on the southern coast in the southeast coast.

All of those could be data center spots as well to Bonnie's point. And, uh, one thing I'll say, number one, Bonnie, you definitely gotta go. It's a fantastic, fantastic country.

Beautiful country. Um, we went there for, for a honeymoon. Um, oh, nice.

Just a great place. Um, Irish people too. Yeah.

Um, but, uh, the west coast of Ireland has no shortage of land, no shortage of wind, um, no shortage of I and maybe can you power data centers on rain and sheep? I don't know. But, uh, maybe make it more sustainable by moving some of this stuff to the west coast on the south coast.

The wind is a great point because Scotland right now, um, has huge, um, initiatives and wind right, uh, wind energy. So I, I agree. I think that that's a great opportunity for them as well.

I just wanna go on the record that Steven said, once again, it's the networking guy's fault. That's all I heard. Definitely.

Yeah. The wind comment was not a slight against Irish people talking, Mike, that was, don't take that the wrong way, that he didn't mean it That way. If, if if talking generated energy we'd be self-sufficient, then there you go.

Well, there is the transatlantic cable that lands in Kark, and that's the, you know, you you got the Yeah, I'm trying to make some kind of varney stone reference. I think, I think the, the issue with the networking stuff though is they want the software to run closer to Europe so that they can take advantage of the axis and the latency. So, you know, Galway may be on the wrong side of the island from their perspective.

Yeah, and it's, I'm sorry, I was just gonna say it's, you know, it's interesting that debate that's becoming more political as, as Mike mentioned, a lot of it has to do with the pressure that they're feeling, um, to reduce carbon emissions and to meet those goals. So it's, it's really like coming from all sides of how they're going to settle this. The moratorium that I mentioned affected a deal that was supposed to happen in August with Google.

So, um, it is already a, a situation where, where not everyone is satisfied. So I think we'll see, uh, some changes ahead next year. Not everybody in Ireland is in love with that, um, climate change regulation because so many of the people out in the west depend upon something known as peat or turf to heat their homes.

And that's now essentially illegal to do because it adds too much to the climate. So guess what's been the latest and greatest thing going on for smugglers? People are throwing Pete in the back of their cars and smuggling it around to each other.

Wow. Black Market. Interesting.

Yeah, it's interesting. I, it, I mean, data is the new oil, right? It, it's, it's, whether it's in transit or we were talking about storage before, and it's interesting how much, not just the economy, but politics now as part of this, it, it almost reminds me of, you know, the, uh, regulatory bodies that control the spectrum that we use for wireless, right?

This is allocated to that. This is allocated to government, this is allocated to that. And we have so much, so many pipes going across the Atlantic and different oceans and, and now points of land are points of presence for these large data interconnection points.

You know, we used to have inter naps, but we do have inter naps in the us right? And now we have the mega versions of that just, I I'm thinking of multiple, you know, blade Runner, other kind of few futuristic movies that, you know, you gotta have data to power all those neon signs in, in the rainy climate that you're running in. Mm-Hmm.

Yeah. And, and the issue will come down in this, the cost of housing in Dublin is through the roof. And those people will say, wait, we're allocating land to data centers.

Mm-hmm. When we need housing, and then the data centers aren't generating jobs, they're gonna be like, you know, ultimately they're gonna say, this is crazy. And they're gonna be like, we need more downstream benefits to the local population before you go and kind of go build the next data center.

And I think, you know, we saw that in this country where, uh, a lot of the politics didn't pay attention to the basic local economics and things changed. And so don't be surprised when that election comes around in Ireland that something different happens. Well, besides a few jobs, and I don't mean to to, you know, undermine what, what the jobs that it may add, but what's the benefit of having that data center there, you know, wherever the location is in the middle of Iowa or the middle of Ireland or whatever location it is, I think if they brought more value than, than just jobs and kind of indirect benefits to the economy, now there's a real reason why you want it there.

Beyond just employing a few additional people, a few thousand extra people, and if they brought value that way to the, to the, you know, to the, uh, economy and to the community, I think there may be a much more of a positive view of having those things located nearby. Mm-Hmm. I agree.

All right, cool. Well folks, we'll be watching this closely because it's gonna play out everywhere in the world, and it will be as election cycles come forward, this is gonna be a bigger, bigger issue all the time. Hey folks, I wanna thank you all for being on the show today, sharing your knowledge and insights.

Great to be here. Great to be here. I'm so glad that we could make the green Pun Too.

It's all about puns at the end of the day. I mean, what's the point of having a show if you can't have puns? We need, We need a good limerick to end the show.

Are you ready? We ready, Mike? Uh, I'm not quite ready for that, but your relatives could do that on demand, Mike.

There's a, there's just, there's a whole county in Ireland if you lack, so you just go visit Limerick and you'll have a great time. All right guys. Thank you all for watching the latest episodes of the Textron Gang.

We have tons of great content coming up right behind this, and till then we'll see them.