Speaker 1: 0:02

Welcome to Uplink, where we explore the world of digital infrastructure, uncovering the technology, fueling AI and cloud innovation, with the leaders making it happen. Well, kevin, thanks for joining us. I know you've got a great background, a lot of experience in our space. I'd love you to quickly touch on your role now, what you've been doing, and a little bit of history, sure, and then I'd love to hear your thoughts on where the industry is going, some of the challenges that you're seeing or opportunities, just how it's all playing out. So let's just take it away. So what's your?

Speaker 2: 0:52

current role Okay, I manage the telecommunications team within NTT Global Data Centers.

Speaker 1: 0:59

Awesome.

Speaker 2: 0:59

Not easy. We're a one-stop shop, so we're involved in everything from due diligence on new builds.

Speaker 1: 1:06

Oh yes.

Speaker 2: 1:07

RFPs for customers, assessing what their connectivity requirements are.

Speaker 1: 1:12

New builds for when you roll out fiber to different locations, or what do you mean by that?

Speaker 2: 1:18

For example, if we, you know, we do our builds by Greenfield builds, right now we don't really do acquisitions. So if we're going to do a new piece of land that we buy, we have to do an assessment of what the fiber and the lay of the land is from a connectivity perspective, so they can have an understanding of how to put that into their…. So when they build the data centers and so forth.

Speaker 1: 1:38

you know that, okay, yeah, so you don't acquire. That's interesting.

Speaker 2: 1:42

No, not yet, yeah well it's cool.

Speaker 1: 1:44

So you're, you're building um. Right, there's a big company yes, ntt lots of different parts of it um japanese uh ownership, I suppose, but everywhere huge, huge and globally it's it's huge, but it's also a bit complicated that it's split up to to sister companies that do like managed services. Okay.

Speaker 2: 2:07

So there is the global data center portion. It's divided between the Americas, eu, india and Asia.

Speaker 1: 2:15

Okay.

Speaker 2: 2:15

But we're in a global consolidation right now, so my director is now going to take on telecommunications role for EU.

Speaker 1: 2:24

We're integrating the EU portion now, so our role is going to expand globallycommunications role for you or integrating the EU portion house, so our roles will expand globally. So US and EU coming together, yeah.

Speaker 2: 2:30

And then eventually India and Asia. But we're in discovery right now because we have to understand what the products are, what the processes are.

Speaker 1: 2:36

Yes, I know that, and before NTT you, I think you were at Equinix for a while.

Speaker 2: 2:41

Yes, I was at Equinix for about 15 years, yeah, so I was there from the fairly early days, so I saw the the extreme growth yes, yeah, crazy growth.

Speaker 1: 2:50

Um, actually, one of the uh, equinix founders sits on our board, jay, uh, yeah, very early, I mean. Well, he found it. Yeah, he found it in the 90s, late 90s, I think it was. Um. Yeah, she told me never be what equinix stands for. I've forgotten now, but it was like it was all the letters, but in internet exchange was the last ix and started with an ix in um I think so hello, yeah, yeah, he told me a bunch of those stories, so, yeah, huge growth there.

Speaker 1: 3:16

What were you involved in from at equinex you said, was it network sides and product sides?

Speaker 2: 3:22

I was in ops, enablement okay so we were kind of the go between the operations side and everything else corporate. So we dealt with a lot of things. So we were heavily involved in product and with Mylene.

Speaker 1: 3:35

Oh yeah, of course she was one of our key people.

Speaker 2: 3:37

we worked with on the product side. So we did a lot with systems development. So any operationally needed changes to the systems we would help lead that with the IT teams and project manage that for the operations side, providing all the SOPs and standardization of policies and training and things of that nature, okay, awesome.

Speaker 1: 3:57

So what are you seeing at the moment? I mean, the world's changing a lot. There's so much opportunity. It's where, when things move fast, particularly infrastructure providers. Infrastructure takes time to build, and so it takes time to figure out where things are going, what they need, and then you've got to physically build something out to do that. So what are you seeing come across your desk? What are the challenges or opportunities, whichever way you want to frame them for you?

Speaker 2: 4:24

Well, from the infrastructure point of view, our interaction is on the design side. We have kind of a model that we do so from connectivity. One of the biggest challenges not challenges, but one of the biggest issues we work with our customers diversity. Right now diversity and redundancy.

Speaker 1: 4:42

Physical diversity so different KMZ paths, so that's different.

Speaker 2: 4:46

Everything Diversity, paths and redundancy on the network.

Speaker 1: 4:50

Yeah, okay.

Speaker 2: 4:51

And so that's what our customers are going over a fine-tooth comb. Yes, you know what kind of whether it's IP or We've experienced challenges with.

Speaker 1: 5:00

I mean, our entire backbone is very diverse. We have to build everything. We ensure that there's separate fiber paths. So we have a KMZ path, which is the actual, you know for listeners like looking at a Google map where the path runs. A lot of cases you can think you have diverse paths, oh yes, but then they just cross in one river crossing or one pipe or whatever it may be duct, and of course that's always the duct that's going to get cut if there's a digger. And all of a sudden you have a multitude of outages or two outages and it and it can be challenging because multiple carriers can potentially share similar ducts. So that.

Speaker 2: 5:38

That's part of what you're asking about. What we do we on the design side, we make sure that we, that the design specs all meet what we do. So right now, our basic design is four zero manholes feeding four entry facilities feeding two MMRs meat mirrors.

Speaker 1: 5:55

So what does that mean? Four separate ingress-egress points to a data center.

Speaker 2: 5:59

Yeah, so the zero manholes the entry point into our property the street, so we'll have four of those Four of those. Feeding into four entry points into the data center.

Speaker 1: 6:09

Wow.

Speaker 2: 6:09

So we'll pull conduit to there and then those will feed into the meet me rooms, which will have two separate meet rooms. So that gives a lot of flexibility for carriers to pathway in with multiple laggards.

Speaker 1: 6:21

So this is the data center itself.

Speaker 2: 6:23

This is the data center itself, and then we'll track which fibers from carriers are going through what conduit. Because if we do, get a customer who's asking for carrier diversity. With these two carriers we need to do exactly what you say Make sure that they're not going out the same conduit.

Speaker 1: 6:37

And do you find that carriers are tearing up and trying to land in multiple, like two sites so that they can deliver diversity to a customer, or is it like we'll put one carrier in this and another carrier on the other?

Speaker 2: 6:50

No, typically our bigger carriers are wanting to have a presence in both meet-me-rooms. Yes, and then can begin, because there's TIA standards about separation that we have to adhere to, because the meet-me-rooms meet that criteria and then they'll pull laterals into Because the mid-room is that criteria and then they'll pull laterals into each mid-room.

Speaker 1: 7:08

So what happens if their network has an issue? So you get all this. This is what I've had come up, all this physical redundancy, even the devices themselves. So if you in our case they're switches, and if you did do maintenance on one, we have a red and a blue. But what happens if the network behind the scenes has an issue? Can that still take it out, or you separate that somehow as well?

Speaker 2: 7:29

Are you talking about the customer's network?

Speaker 1: 7:31

The carriers more the carriers that are providing those components.

Speaker 2: 7:35

Well, that's the reason we're going with diversity.

Speaker 1: 7:39

So whatever does have them down a bit.

Speaker 2: 7:42

Or even diverse pathways, because the diversity requirements comes from our customers, yes, so we provide options and solutions that meet what their requirements are. And each customer kind of has their own definition of what diversity means to them. Yes, yeah, because you can say that and it can have a totally different meaning to another customer. So we kind of have to dig and ask questions of what their tolerance risk tolerance, and then we'll meet that. So it can be you know, being a carrier neutral data center.

Speaker 1: 8:15

We have quite a few carriers, especially our retail areas, and then we can satisfy most everything we I mean we've got interesting designs where we've seen customers they want two physical data center locations in case the entire data center has an issue.

Speaker 1: 8:32

So then you have and then you ensure they're on different grids or even different locations. You could even say like one east coast, us, one west coast, and then you could even go even country. But we've seen Examples were on the East Coast. At West Coast you have the diversity of physical data centers and then inside those data centers we have two physical devices which are totally separated physically and we treat them as a red and a blue zone so they could never be upgraded At the same time. They're treated separately.

Speaker 1: 9:02

I know I've seen it, yeah, and then on top of that we always take two. In some countries we take three or four or up to six paths, potentially depending on the country, but in the US it's usually two physical diversity paths, so that every time our customer and this is one of the values we bring is that any VXC or virtual connection from Megaport, you are instantly resilient across the physical infrastructure and so you could take two ports. That ensures you're physically connecting over red and blue, but the path that it takes is always running across multiple. So instead of having to buy one path and then another path from a separate area, megaport's already done that work for you and virtually connecting across that, but we've also then seen them take another data center in case something happened to that data center. It's all about the risk tolerance.

Speaker 1: 9:57

Yeah.

Speaker 2: 9:58

And it's kind of a standard model we'll have. Where they'll have a deployment, then they'll have a data recovery deployment across country. Both of those will be tied into the cloud. Yes, a lot of times we're using megaport for that, and then even on the cloud side, because multi-cloud is now becoming the big thing, the big issue. So but, they may run the a, you know the red router to a red, you know to a on-ramp in one data center and run the blue to a different on-ramp and another data center on the cloud side.

Speaker 1: 10:25

Yeah, we see that a lot. Yeah, if the if the on-ramps allow another data center on the cloud side.

Speaker 2: 10:28

So we see that a lot, If the on-ramps allow that. Sometimes the locations don't work for that.

Speaker 1: 10:32

We've seen Azure has done a pretty good job of putting an SLA around connecting into the clouds, and so we've actually we've seen a lot of customers, I think because of the push, because they're getting SLA, with diversity, having to ensure the different paths. I think it's a great option because it's actually sort of being pushed. I think some of the other clouds hadn't pushed that SLA component so that weren't as focused, but we know that that makes a big difference.

Speaker 2: 10:59

So the network you were talking about and the redundancy you're talking about, we go over that with our customers quite a bit because they ask and we kind of get them. Sometimes we'll bring your sales engineers on to explain that for what?

Speaker 1: 11:13

totally. Oh, that's very. It works really well. We've had good response from customers. So, resiliency design what are you seeing in terms of demand? Is ai changing? Anything is? I have much to talk about that, but what are you seeing this change? Hyperscalers I don't know what's the well, what's the trending you're seeing in your world?

Speaker 2: 11:28

where you know. Wave one was the internet, yeah, you know the data center. Wave two is kind of the uh, the cloud. Yes, and now we're into a wave three, the ai side, and it started to change around the beginning of it. You know from an internal. It's changing just how the data center works in terms of data centers, using the AI to make the data center run more efficiently.

Speaker 1: 11:50

So from an internal use, that's obviously one.

Speaker 2: 11:52

But from a customer perspective, it's not so much the connectivity side's bandwidth, it's driving higher bandwidth requirements. So you know the old yeah. I can remember back in the day I'm sure you do too when OC192 was considered the biggest pipe in Israel. Every, every, every yeah, and. But now we're talking about 100 gig becoming almost a basic order yes, yeah 400 gig now is what they're looking for. Yeah so that's coming up, and so bandwidth is one big change, and I think AI is driving that heavily yeah.

Speaker 1: 12:36

The second big change is not so much on the connectivity side but the cooling you know the traditional yeah, um, and this is this.

Speaker 2: 12:40

I mean liquid cooling, water cooling and so forth. Oh, how's that? Yeah, sorry, yes, so liquid cooling, we do both. We do immersion cooling as well. What do you see?

Speaker 1: 12:45

um what do you seeing customers adopt Both?

Speaker 2: 12:48

Both yeah, Again.

Speaker 1: 12:50

Depending.

Speaker 2: 12:50

Yeah, again being a co-we rely on what the customer's requirements are. And they're the ones kind of telling us we have to just make sure that the data centers are ready to accommodate those kinds of cooling techniques. Yeah, I have an interesting and a little bit of side my wife's in genetics, oh yeah. Yeah, I read some of her articles and there was an article I was reading that in san francisco some of the bioengineering firms are looking at creating genetically modified organisms that will pull heat out of the cooling systems and they they see that as a long-term solution to ai.

Speaker 2: 13:26

So just shows you never know where technology is coming from.

Speaker 1: 13:30

Yeah, it certainly helps solve the carbon emissions related to the energy related to it. Yeah, that's cool. There's so much cool stuff coming. So do you feel let me ask you this one You've been in the industry a while it always feels whenever I, if I went to any moment in my career and said, went back and talked to myself, I would convince you that it's the most disruptive moment in history, and it feels like it's always more disruptive. Now it feels more disruptive now that I did, you know, three years ago, three years and back and forth. And so the question is do you think it is always just disruptive and always something changing? Or right now it is moving faster than we've ever moved before and that is then changing the game, or is it? Or there's some days you wake up and you just go. You know what. It's the same thing that we've been doing for a long period of time. You know what I mean.

Speaker 2: 14:19

It might be just faster I think it's a combination of both I think it's a combination of both I mean we are doing a lot of the same things in connectivity that the core concepts haven't changed that much but the speed at which technology is progressing makes it feel like you're just never ahead of the game you can never catch, and once you learn something, you just you. You hear a conversation from somebody of something you didn't know and it just puts a light bulb on. Hey, we can apply that back to here.

Speaker 1: 14:49

Yeah.

Speaker 2: 14:50

So I think it's a combination of both. Yeah, the four things are the same you still have to connect, you still have to use the internet, you still have to do those things. It's just that how you do it. I mean a good example you're ahead of the game. You're pushing a lot of our other carriers to start creating a lot more flexibility on their ordering, so we're already pressing our big tier ones to say, hey well, I want to have a portal where I can order a wave, submit my things in and get it in a couple of days and not have to wait a month and go through all this.

Speaker 1: 15:21

And spin it up and spin it down. You know what's so fascinating on our side? We're 11 years old, uh, and we're still really disruptive. And so let's say, some things change fast, right, really fast, and some things change really slow. I mean, you would have thought that what we brought out 11 years ago, everyone would just have caught on and done that. Now the interesting part is and this is what I think about it a lot, because our, our shareholders would ask why aren't there more competitors in your space? Why aren't there more people doing what you're doing? Now, it's not because they don't want to, they absolutely want to.

Speaker 1: 15:53

But the key to what we deliver is two things. One, you have to automate it, because otherwise it can't be delivered quickly. You can't it's humans take too long. It can't be delivered quickly, you can't, it's humans take too long. Now, if you automate something and you do it in less than a minute, it means you can go spin it up, spin it down In every five minutes. You could make big changes to the speed. So you have to have automation down to less than a minute at least.

Speaker 1: 16:15

But the second complex part is the billing platforms. Don't underestimate how complex they are for most of these telcos, particularly when they've got multiple acquisitions. Many of them don't have one billion platform. They've literally just got a billion platform for every acquisition. They couldn't link them together. So we're finding the limitations. It's not like technically what could be done, it's the machines behind them, and that's why we have this luxury of building from ground up automation. Every single piece of equipment was chosen and automated from the get-go. I don't know how you do it the other way around. We've seen that it's not really happening. So as much as you push, I think that that's why we still see ourselves as this really differentiated thing, even though it's 11 years old. So some tech super fast, some a little bit slower, so we're sort of the beneficiary of a slower industry on that side. But um, we're keeping up with this insatiable change on everything else.

Speaker 2: 17:15

So what you're talking about isn't a technical issue. It's really the change management behind it. Yes, and use those technologies correct. It's a challenge, especially with bigger companies that have done a lot of acquisition. Yes, absolutely correct.

Speaker 1: 17:27

Well, the ones that should automate. Now, the other thing that's also interesting is you talked about resilience and redundancy. One of the things that a carrier can deliver and they'll talk about their strength is their fiber assets, as a great example. Our strength is that we have no fiber assets, because we can pick and choose and always ensure different kms is totally different networks. We can take completely different paths, as opposed to one provider only using their own infrastructure, and that is the key for us to ensure that our customers have full diversity as well. So it's almost like a strength by not having the infrastructure, whereas they would say the strength is having it. So we're just seeing this interesting space play out as well.

Speaker 2: 18:10

Yeah, yeah, so but we're happy with it. I mean, mayport's been a really good partner for ours and it's offered us a good alternative to some of the wave connections we've experienced In the US. The fiber there are areas where there's still a lot of aerial fiber, especially in older areas where they have to go to water and stuff like that. And they're really prone to accidents and stuff. So when you have a resilient or redundant network like yours, it really offers us another solution in case the waves connections just aren't working.

Speaker 1: 18:45

The pathways don't work for diverse pathways yeah, I mean we just there was an outage, um, spain, portugal, um, uh, that had, there was a big outage with power actually just recently. Yeah, we had this and we didn't even know about it. Um, because the network constantly self-heals this way it's fixing. There'll be something wrong, the team will be fixing it, but from a user, they don't even notice it. And the only reason I found out about it was that customer sent us an email thanking us for being the only one that remained up through all this time and fully available. And I was like what, what was the outage? And then you realize I think there was some power outage over there.

Speaker 1: 19:23

Anyway, the purpose, what we've built is built expecting failure. We expect someone to dig a hole, we expect someone to drive through a thing and rip out a bit of fiber or whatever, and if you expect that you build another path, you build a resilient for it. So, yeah, at any moment in time, we've got different outages, but it's never actually impacting the network as a whole. So, yeah, super important. Well, we appreciate your partnership. What about, um? What about the future? What can you say? It's just more, more of the same, or does it keep getting disrupted, or I don't know. What are you thinking from a design side?

Speaker 2: 20:02

that's a tough one to play to be honest, we're so focused on trying to meet the changing environment. Right now I don't really have that. Have really thought it up. Um, I think ai is going to be the driver and there's a lot of unknowns. I mean, you have things like quantum computing now starting to become a reality, especially if you're dealing in any of the crypto or cryptology solutions. It's just bandwidth, you know.

Speaker 1: 20:35

That's why I think.

Speaker 2: 20:36

For us. I think it's more the same for right now.

Speaker 1: 20:38

Yeah, I think about that as well. We're like, particularly as a CEO of a company, we've got to place bets, investments, in the future. Um, if you go back, you know, 20 years, it's still the same. We're still connecting. What's happening is more resiliency, more speed, more power, more locations. All it's just more right. But you can hang your hat on the fact that networking actually hasn't changed a great deal in that time. Still, ip addresses, a lot of IPv4 still out there, the structures of how actual routing works and how you fail over or don't fail over or you flap, and all this. It's actually not changing that much. So that's nice to be in an industry where we'll continue to be critical for what is being rolled out, regardless of whichever path it rolls, and we'll continue to support it. It's just yes, to some extent you don't have to place too big a bet, I think on that side.

Speaker 2: 21:30

I think, thinking a little bit about what's happening. I see a lot of the smaller customers moving out of data centers and going to the cloud directly from their corporate side. So they're kind of abandoning the private cloud model. Now the enterprise is still doing that, but that's one thing. I see more, so you're seeing it.

Speaker 1: 21:49

You'll see small end moving to yeah.

Speaker 2: 21:52

Kind of going up, and so that's probably one thing I can see. It's a trend right now in our data centers.

Speaker 1: 22:02

Yeah, well, we appreciate the partnership. Anything you need from us, you let us know. Otherwise, yeah, keep building what you're building and we'll service these customers together and make an impact.

Speaker 2: 22:12

Yeah, your teams have been very helpful. So I just want to say that not only the account teams but also the sales engineer teams have been really invaluable, because when we're going into the cloud, a lot of times we don't have the cloud experience. Our customers don't always know how to interact with the Megaport coming into the cloud and your sales engineers are always helping us out on that side we're learning a bit. It's a good partnership. We're really happy with it.

Speaker 1: 22:37

That's great. Our solutions architects are so critical. What's super difficult is each cloud provider has a different way of doing things. How you build resiliency, how you design for egress, will change your costs dramatically. We launched a cool new product called NAT Gateway, which is saving customers a huge amount of costs just for NAT translation. We're seeing customers use that and then push that back into the cloud and use that spend for AI and all sorts of things. So, yeah, we appreciate it. Thank you, thanks, cheers, yeah, awesome Thanks.