AI Is Exposing Bad Network Decisions

0:00
A lot of times enterprises stick with what they've had for years or
0:06
what the last CI CTO CIO said to put in there. So what we find is we say um why
0:13
they say well we have waves connecting these two data centers and this is the
0:16
workload that we do and we go okay well why do you have that wave because it's
0:20
really long or it's not performing and there's a reason. And so they say well
0:23
that's kind of what we've always had. Welcome to Signed, the podcast for
0:28
buyers in a market built for sellers. I'm Max Clark, CEO of ITbroker.com.
0:32
Today I'm sitting down with Scott Nichols, chief commercial officer at
0:35
Aurelion. Let's go. I still very much like talking about
0:47
technology and what's going on with technology, but what I found was
0:52
>> the more time we spent there, the more it was more about how are we applying
0:55
technology to solve a business need, right? And and so then, you know, while
1:00
I'm still interested in, you know, still very much like a network engineer at
1:03
heart and like still have that kind of, you know, mindset around things and
1:07
wanting to design and build, it shifts a little bit. It was like, "Okay, great."
1:11
You know, so you've got this thing here. Why do people actually buy it? You know,
1:14
how do you help somebody through the buying journey of this stuff? What's the
1:17
evaluation process really like? How does this thing get evolved? And so, like,
1:21
you know, so you ask like list of questions.
1:24
>> I just keep this because, you know, it's very easy for me to want
1:28
to like wander off into the weeds sometimes.
1:30
>> Guidance after 200, you kind of have a
1:32
>> Yeah. You know, see, you just have you just kind of have a thing where you're
1:34
like, "Okay, you know, are we kind of" and this
1:37
>> um,
1:38
>> you know, I I I wrote these out, you know, five, six years ago. This
1:42
grounding um, you know, experience of like, you
1:47
know, can I go back? Are we are we, you know, probably the best advice I've ever
1:51
seen from anybody was like, if you write out your intentions before you do
1:54
something, you can always go back to your what you what you had written out
1:56
and it's like, are you adhering to your intentions? And, you know, so I like
1:59
that, you know.
2:00
>> No, it makes perfect sense. Well, so to to I'll get on something you said. So I
2:04
am not an engineer. I am I'm I started out as a a telemarketer back in the day
2:10
of when telemarketing was a thing for Sprint and I've always been in sales.
2:15
I've only sold technology. So uh you know I'm chief commercial officer at
2:20
Aurelion. So I've got sales, sales operations, uh marketing and one small
2:25
piece of procurement called quoting and ordering. But I know enough to be
2:28
dangerous. But to say that I'm an engineer and I just if you go way deep
2:32
in technology, I'm gonna I'm not gonna be able to answer.
2:34
>> I'm not gonna ask you about BGP communities.
2:36
>> Thank you. Don't don't worry. We won't get into that part.
2:40
>> Which is definitely a conversation that we have with our clients all the time,
2:42
but I say, "Let me get you with an engineer to talk about that."
2:45
>> You know, the conversation with your client about BGP communities is like,
2:48
"Give me every tool available to me as a widget so that way I can do my own
2:52
stuff." Right? Like that's not you know like it's um
2:58
Steve Jobs is famous, you know, if you ask people in a focus group what do they
3:00
want? They're like they just want more of what you're already giving them. So
3:03
you know once you say you say you have this you can do traffic engineering you
3:07
know with with metrics and communities and and whatever else are just like yeah
3:11
of course more you know give me all of them you know so it's
3:14
>> well I come from primarily from the enterprise space so I've lived in that
3:17
space. I worked at Ma G and uh um Xperio for a bit and then at Aurelion, you
3:23
know, we've been a 30-year lineage in being one of the largest wholesale
3:26
companies in the world. And so about three years ago, 5 years ago, we said
3:31
we'll make the move into enterprise. And at the end of the day, um we sell the
3:36
same products, right? We're very we're very focused on connectivity. We don't
3:40
do SDWAN. We don't do voice. We try to stay focused on what we're good at. But
3:45
what's interesting about it and what I see is that the the enterprise market
3:50
layer in AI AI and cloud and all the things that are going from sight to
3:55
sight to data center to data center cloud instance to data center the market
4:00
is kind of rising to meet us in that in the enterprise space I'm having more
4:04
conversations about IP transit than we've ever had before. We're seeing a
4:07
rebirth of Ethernet, believe it or not, in a certain way because it still is one
4:13
of the best performing technologies out there. Most of my competitors have like
4:16
a separate Ethernet network that's oversubscribed and not working. Ours is
4:20
all in one. It's designed for that. So, my Ethernet performs better than most.
4:24
And so, people are kind of like, if you can be price competitive, I'm
4:27
interested. It's just that I've always been priced out versus a Wave or a U,
4:32
you know, DIA. So anyway, I'm seeing some interesting things coming from the
4:37
enterprise space, but living in the wholesale lineage of Aurelion, formerly
4:41
Telia, but the enterprise is it it it's it's changed a lot. So I mean, and
4:46
there's not much more I can say technically. I think I've expanded. No,
4:50
I'm kidding. But you know, that's my level of conversation.
4:53
>> Let's let's start and talk about wholesale,
4:56
>> you know, carrier, right?
4:58
>> And um pre200, right? you know, let's say like the real the real acceleration
5:03
of what we see as the cloud today, you know, the AWS, the the Google, you know,
5:07
Oracle, whatnot. Um, it was really common. We had an ecosystem where uh
5:13
companies had data center builds, 10 cabinets, 20 cabinets, you know, much
5:18
smaller footprints typically in a data center at that point, right? So maybe
5:21
100 to 250 kilowatt in a data center. and and there was a a relatively robust
5:26
ecosystem of tier 2 networks, you know, networks that were buying from tier one,
5:30
you know, from telly at the time, right? you know and your competitors and uh and
5:35
then going and providing layering service on top of that and giving to the
5:38
enterprise and whether that enterprise was you know true enterprise when I'd
5:44
say like true enterprise I mean like you're you're making revenue that's not
5:47
connected to the internet or something or you know if you're in that like e e
5:51
company right the SAS companies
5:53
>> that was pretty typical and we saw you know IP transit and IP transport being
5:58
dominant you know structures at that point now AWS US we'll just say cloud
6:04
cloud comes out and now you've got interesting friction right because the
6:08
cloud companies become your customers
6:10
>> but at the same time very much your competitor for the end user right and
6:16
then enterprise start shifting up into the cloud
6:19
>> cloud fabric thing they sell
6:21
>> so how you know and so now you know everything is circular and tech
6:28
right so now we see enterprises are kind of coming back to do we need data
6:31
centers. Do we want to have our own physical infrastructure? Do we want to
6:34
have committed capacity somewhere? Do we have performance advantages or financial
6:37
advantages or you know a whole slew of different things come
6:42
into this equation that wasn't just this wholesale like put everything here now
6:45
it's it'll be a little more temperate as you're as you're looking at that and
6:51
you start talking about you know engaging with an enterprise
6:55
now now this enterprise has workloads in lots of different places
6:58
>> and how is that tension and friction working out with your network and and
7:05
your services you're delivering. Well, I mean, for us, you know, a lot of times
7:11
enterprises stick with what they've had for years or
7:15
what the last CI CTO CIO said to put in there. So, what we find is we say, um,
7:22
why they say, well, we have waves connecting these two data centers and
7:25
this is the workload that we do. And we go, okay, but why do you have that wave?
7:29
Because it's really long or or or we get, you know, it's not performing, and
7:32
there's a reason. And so they say, "Well, that's kind of what we've always
7:36
had." And we said, "Well, let's look at your bigger network picture. We could
7:39
connect that with EVPL." And then, you know, you you you can still have some
7:43
path diversity, but you it won't be nailed down,
7:48
>> but the the the performance will be better because with our EVPL, we can,
7:53
you know, if there's a link that goes down, we can reroute that. So you might
7:57
have a little latency, but it's not going to impact your your workload or
8:00
your user experience because right now your wave is up or it's down.
8:04
>> And and so it's that's an example of a conversation where we just say why what
8:08
are you using it for and how are you using it? And so for us, like I said,
8:11
we're focused just on connectivity. So it's I I I sell five things, you know,
8:16
so it's a very different conversation and I can really drill down. I've got
8:19
great engineers. I've got great, you know, network planners and and um uh
8:24
network um uh you know, uh I'm trying to think of the the role I can see the
8:29
guy's face, but there's a lot of different things that we can do because
8:33
we're so focused just on this. And so the conversations are interesting
8:37
because we're challenging some of the the norms on what they've used. But all
8:42
you know when you look at the enterprise landscape a lot of folks went and they
8:46
said we need SDWAN for you know application resiliency and and um you
8:51
know blackout brownouts and all the great things that SDWAN gives you. And
8:55
that's awesome but a lot of times SDWAN just told you that you've made terrible
8:58
decisions on your underlay. And so you know we come in and say well
9:03
why did you choose this and it was somebody else made a cost decision. I
9:06
thought I could get away from Ethernet and DIA and just go DIA and fixed
9:09
wireless or DIA and broadband. Those are great solutions, but there's the right
9:13
solution depending on the type of site and the type of traffic and the type of
9:17
application environment, you know.
9:18
>> So, we say, look, you got an A, B, and C site. Let's go A site. Let's go let's go
9:23
Ethernet with DIA. And I don't care what you do BC, but let's have the
9:26
conversation about what you're using it for and why. We're seeing more and more
9:30
of that. But what's constant in the enterprise space is the bandwidth
9:36
requirement is doing nothing but grow. So SDWAN making changes becoming
9:41
commoditized voice very commoditized and we're not in those businesses for that
9:46
reason but internet cost per meg's always going down there's always price
9:50
pressure but what you're seeing is the consumption requirement and the need is
9:54
only increasing and performance is increasing as well so it's opening up
9:58
other technologies that people said oh you know Ethernet's dead I don't I it's
10:02
definitely declining but the use case is still there but you have to talk about
10:08
Does that make sense?
10:08
>> The the whole Ethernet is dead thing is so funny to me. It reminds me of like
10:11
early 2000s when an article would be, you know, published of like the
10:14
internet's running out of bandwidth and you're like, what are you talking about?
10:17
Yeah. You know, you have no idea what you're talking about. And you you know,
10:23
my mind churns because as you're as you're mentioning different things, you
10:26
know, I go lots of different places technically, right? you know, you go
10:29
into the whole thing of like, you know, when you're selling EVPL
10:34
over, you know, versus NLS in most cases, most networks or even just an IP
10:39
circuit on top of it. In most cases, most service provider networks is the
10:43
same infrastructure. So, trying to have that conversation with somebody and
10:46
saying, why are you buying NS here versus just using internet? It's got the
10:50
same routing path between point A and point B. Like, what's your objective
10:53
that you're trying to achieve with this? Um, you know, it was it was common for
10:58
networks for a while to have an optical core and then have some sort of IP core
11:01
on top of it. So then we can get into an argument about whether or not buying
11:03
waves is more efficient on that network or not. But even even that isn't the
11:07
same consideration in most places anymore. Like it's it's it's evolved and
11:11
changed and SDWAN I I'm I'm laughing because you know we get into SDWAN
11:16
conversations and somebody says we want SDWAN. I'm like okay great. Which kind
11:20
of SDWAN? Right. Right.
11:22
>> Because there's like five, right? you know, which problem are you actually
11:24
solving?
11:25
>> That's it's the same conversation. Why? What do you need it for?
11:27
>> Yeah.
11:27
>> So,
11:28
>> a lot of people are just like, we can't figure out how to configure VPN on our
11:31
on our firewall and you know, the SD1 just solves that problem for us. It just
11:34
creates this magic, you know, VPN overlay. And I'm like, okay, yeah. I
11:38
coming from somebody who had to configure a VPN on on Cisco routers and,
11:42
you know, picks firewalls. I like ASAs. I'm like, I totally get it. You know,
11:45
you configure your first one, you're like, it works. I don't know why it's
11:47
working. Stop breathing. You know, like it's like if it ever goes down, I don't
11:50
know why. Well, SDWAN isn't anything new. It's packaging of some existing
11:54
technologies in a different way, but you know what we're seeing, you know, we
11:58
thought about it when I came on four years ago. They're like, do you know,
12:01
we're going to go after the enterprise market in a bigger way. We're going to
12:03
focus on, you know, very subset of a subset. Do you want to sell SDWAN? I'm
12:07
like, no, I don't think we do
12:09
>> because nobody's selling enough of it to make it a huge sustainable part of their
12:13
business. And then it's a very messy marketplace. like you said there's
12:18
different different depending on what you're trying to solve you you buy
12:22
different devices stacks of software etc. the tension I see with SDWAN um the
12:28
people you know I think a misconception that I deal with a lot is understanding
12:31
the economics and the cost structure behind it and so you have SDWAN vendors
12:35
and even now I mean Sassy is a little different animal because you're getting
12:37
a security right that's
12:39
>> but you know SDWAN vendors that are that are you know um pinning traffic against
12:44
some sort of infrastructure and a lot of those people are pinning traffic against
12:47
some sort of cloud infrastructure you know using one of the big cloud
12:50
providers and all of a sudden you're like why is your cost structure so
12:53
upside down you're like oh I get why your cost structure go upside down
12:55
because you've got this underlying cost that you can't deal with, right? Like
12:59
it's it's um you know and then you know it's it's um that whole dynamic is
13:06
really
13:08
>> it it it feels like um so many of what we see in terms of SD1 being deployed.
13:13
It's just we don't we don't have networking talent that actually can help
13:17
us design and architect and maintain or we're we're we're trying to solve the
13:20
evils of some OEM that we've already installed that we're committed to that
13:24
we can't fix so we can put something else here and kind of solve that
13:26
problem.
13:28
>> Enterprise um I'm curious about this. My I'm a big
13:33
advocate for every enterprise to go out and find and and contract a small amount
13:39
of data center space near their operations, especially their large sites
13:44
because it's like such a gateway to everything you could ever want to do.
13:47
Whether you're purchasing bandwidth for a lot, you know, you have completely
13:50
different economics of purchasing bandwidth, you know, any sort of
13:53
transport changes, cloud ons change, everything changes when you're in a
13:56
carry neutral facility close to you. And then metro circuits of course
14:00
economically change. Yeah,
14:03
>> Aurelion isn't a last mile provider. Correct. You're not in the United States
14:07
like as a cable company or an eyeball network. So,
14:12
how how is that dynamic been playing out for you with enterprises that start to
14:16
like make this understanding and switch of like I don't want to use my local
14:19
lack or I don't want to use a local MSO because of whatever reason, capacity,
14:25
cost, you know, reach, whatnot. um you know that makes interesting
14:31
waters for you to navigate.
14:33
>> Yes and no. I mean it's something we've been doing for a long time right in the
14:36
in the wholesale space because you don't talk about last mile when you're talking
14:40
to you know carriers. But as as as enterprises shift to your point more to
14:44
the data center model cloud as well uh connectivity conversations change it's
14:49
not so much sight to size it as it is you know data center data center data
14:53
center cloud. Um, we still see quite a bit of conversations around I still need
15:01
a tail circuit. Um, when we go in, we'd love to just talk about IP transit and
15:07
waves and and Ethernet, right? But that's not where the problems lie. So,
15:12
when you meet with an enterprise, you say, you know, what can I help you solve
15:15
achieve, right? That's really the conversation. And a lot of times it is,
15:20
I've made some, you know, commitments to this carrier. they use this tail
15:24
circuit. I don't I don't know who it is. You know, a lot of times they don't know
15:27
who it is. They just know it's not working. So, we'd say, "Okay, well,
15:31
that's a great conversation, but I'd really just like to talk about the stuff
15:35
that's working well, which is your waves, your IP, you know, and and build
15:38
on that, but they go, you got to solve my problem first."
15:41
>> That's reasonable. So, we still have quite a bit of conversations around
15:45
physical diversity, uh, pop diversity when it comes to tail circuits. So even
15:50
though the dynamic has changed, it's it's still part of our conversation and
15:55
you know most of the carriers in the US as well as the large ones and small ones
15:59
globally are customers of ours.
16:01
>> So buying tail circuits and deciding on what's the best path, how do we make
16:07
diverse from whatever they're trying to at a site loca location based isn't
16:12
anything unique to us. We've been buying tail circuits for 30 years. We just do a
16:16
lot more of it now because it's part of the enterprise need. Does that make
16:19
sense?
16:20
>> Of course. Marketing, right, with networking
16:24
companies or or carriers. Wireless companies have been marketing 5G for it
16:29
feels like forever at this point. Y and a lot of that marketing has been like
16:34
I I think um trying to invent a problem or a need within it. Oh, you know, we're
16:41
going to provide fill in the blank, right? you know, autonomy because you
16:45
can do low latency and sub whatever blah blah blah blah blah to the edge and and
16:50
it's been, you know, in my shoes, it's it's it's always like we're trying to
16:54
invent the problem that people actually need to solve. And you're like, nobody's
16:57
actually trying to solve that problem because for most applications, if you're
17:00
talking about inside of a metro major, you know, market, you're already really
17:03
fast. I mean, latency, you know, what's the application that you need sub two
17:07
milliseconds of latency for, right? Like doesn't really exist. And with AI in
17:13
every segment of tech, everybody's running in with like what our AI
17:17
strategy is and we're an AI company. We're an AI network this that, you know,
17:21
like all these really think
17:23
>> and on one hand it still it still feels like a similar case of smok and mirrors
17:28
and like hyperactive marketing trying to get onto the latest wave. But on the
17:31
other side, we see really interesting demand patterns and curves changing, you
17:38
know, and the amount of data that's flowing back and forth in these things.
17:41
So, how has AI really change this demand curve for you with enterprises and the
17:45
conversations you're having? So, you know, you mentioned wireless. I I
17:50
definitely think there's a place for it. I think it's really cool technology. I
17:52
spent a lot of time in that space, so I keep up with it a bit. But, you know,
17:55
I'm I'm a fan as a user and I just think it's really interesting and there's
17:58
definitely a place for it in the enterprise sort of, you know, landscape.
18:02
But, you know, AI for us, you know, you can get very technical and dig into, you
18:08
know, GPUs and all different kinds of technologies. But for us, what we see
18:12
regardless of whether you're a platform provider, u you're an infrastructure
18:16
provider of AI, you're a platform provider using leveraging AI or you're
18:20
an enterprise leveraging an AI application, if you will. It just drives
18:25
more traffic.
18:26
>> I mean, simply put, it just brings so much more traffic to the game. And and
18:31
as I mentioned before, we're seeing that as a big driver for somebody talking to
18:35
a company like a Railand because, you know, you look at Kentech, who's got the
18:39
biggest global backbone in the world. It's us. Um, we were the first to launch
18:44
400 gig, uh, Waves, IP Transit, and we will be launching Ethernet fairly soon.
18:50
Um, we'll probably be the first to launch 800 gig. So our we just passed a
18:55
threshold I think it was February where we did we 200 uh terabits per second
19:00
across our backbone. So if you're wanting to say where do I go to the
19:06
company that has the best you know who can accommodate my growing needs and my
19:10
traffic at the top of the list. So that's really what's getting us in the
19:13
door because they're seeing more traffic from AI than they've ever seen before
19:18
and and it's not stopping because you everybody has what's our strategy and
19:22
the answer is we're developing it a lot. You know they kind of make it up as they
19:25
go because it's just so new to the enterprise space. So, so it it's one of
19:30
those things where I think that AI is really driving things where our
19:34
strengths, the market's kind of rising to to meet us because, you know, not too
19:40
many companies can do 400 gig waves, 400 gig IP transit and talk about 800,
19:45
>> right? So, it puts us in a good spot. Now, we're not a fit for everybody, but
19:48
at that at at the top top end of the market across multiple verticals where
19:53
they're seeing a lot of use cases, then we're in those conversations. So again,
19:57
it's it's a well, I'd never heard of you guys before. You know, people have heard
20:01
of Telia before, right? But but we've been in business for over 30 years.
20:06
We've been in the US for 27 years. We've got 3400 POPs globally. We're, like I
20:11
said, we're the only one that has a a one of the few that have a global
20:15
backbone. We can do both DWDM systems and IP Transit. We we maintain both
20:19
across the same backbone. So there's a lot of things that we do that a lot a
20:23
lot of companies don't do and they just haven't heard the name.
20:27
You know, if you were involved in actual backbone engineering, you know, right?
20:30
Yes. But if you
20:32
>> if you weren't, you don't, right? You know, it's just it's um
20:35
>> especially in the enterprise, if you advertised an AS number, you know 1229
20:39
>> and and so you're very familiar with us. You may not know the name of the
20:41
company, but you know the the AS number. But if you're an enterprise, different
20:46
story.
20:49
>> I'm I wouldn't I'm not going to use the word term surprised. Um, but I'm still
20:53
fascinated to see how like like this skills gap enterprises have and um and I
21:01
and I credit I mean SDWAN solved a problem which was you know how do you
21:05
deal with network engineering and skills gaps with network engineering but then
21:08
at the same time it promotes an even bigger skills gap because people aren't
21:11
coming up that actually understand like how do you do this? How do you
21:13
architect? How do you design? How do you how do you configure BGP? Like you know
21:17
you you still see stuff all the time on the internet. It's like, oh, you know,
21:21
somebody's doing something they shouldn't have with the routing table.
21:23
And like, you know, there's like a little thing that happens. Um,
21:28
from, you know, from from your vantage point,
21:31
right, you have to deliver a service, but you also have to create customers
21:35
that can actually consume the service. And, you know, how do we how do we
21:39
create and and, you know, establish like the next generation of network engineers
21:44
and actually understand, you know, how do you how do you plan a backbone at at
21:48
a large scale? like how do you make this decision, right?
21:51
>> That's a that's a tough question. I mean, the the the
21:55
the challenge I think all carriers have is the aging population and the lack of
22:02
knowledge like you said in the space today.
22:05
>> Young people coming out of college, it's not sexy to work at a carrier like it is
22:09
at an AI company or pick one. So um for us you know most of our solution
22:15
engineers are the secret sauce the the network planners and the IP engineers
22:20
are the secret sauce. So we we bring a team approach. And so I think that's
22:24
what makes us a bit more sticky is that when you when you are trying to change
22:29
how you're set up your network or you have questions or you've done it wrong,
22:34
that's really when we bring in the the big guns that are behind our global
22:38
backbone that understand how this works on a global scale and we say look let's
22:42
talk about how this works and come up. We we have multiple iterations on the
22:46
architecture of the solutions and all those kind of things. To me, that's all
22:50
positive. Um, those questions and those workshops that we put together are huge
22:55
buying signs and it becomes, you know, you're a partner in the business, not
22:59
just a vendor. But we definitely leverage our secret sauce, which is our
23:03
solution engineers and our IP planners and backbone, you know, and uh and um IP
23:08
provisioners and really the guys know the peering managers. That's what I was
23:11
looking for earlier, the peering managers and how all this works because
23:14
it is sort of uncharted territory for some of these folks and there is a gap
23:17
there. But we bridge that gap because this is all we've done for 30 years.
23:21
I've not had to be an SDWAN expert. We've not tried to be a voice expert or
23:25
a security expert. We've stayed focused on connectivity.
23:30
Peering is an interesting conversation kind of move into for a second, right?
23:33
Because this is this is a capacity planning and and you know you're scaling
23:37
400 gig 800 gig which means your customers are scaling to 400 gig, right?
23:41
Yep. And you know where one gig was common replaces with 10 gig now like 100
23:45
gig is common and people like you know the four early adapters are moving into
23:49
the faster and faster speeds but now you've got the leagards. The leards are
23:53
typically the eyeball networks right or certain
23:57
you know geographies let's say like Asian countries where you have really
24:01
fast eyeball networks that are coming up but then also interesting politics.
24:05
>> Yes.
24:06
>> Um I'm not going to name any country by name here. we'll keep this friendly but
24:10
you have interesting politics that then come into peering in region peering out
24:13
of region peering you know what what capacity is transiting an ocean or not
24:18
and that creates pressure on you as well because you know you have to kind of
24:24
increase I mean saying we're going to go from 400 gig to 800 gig isn't just like
24:27
we're going to change some line cards out right now you have to look at the
24:30
whole system as a whole and be like how do we elevate the entire thing right and
24:35
what's what's also been wild to w you know witness over the years is like Oh,
24:39
we just we just turned up a new peering link in capacity with whatever network
24:43
and like it was turned on and then it went to 100% capacity instantaneously,
24:47
you know, and you know, so how you know like you know years ago we had like the
24:52
peering battles and the peering wars and the peering cakes being sent back and
24:55
forth and we haven't really seen this stuff for a little while. it seems to
24:57
have simmered down a little bit but um how how are you seeing this with the
25:02
eyeball networks you know whether it's you know in like the United States or
25:07
internationally kind of at the same time competing with rising bandwidth demands
25:12
and you know how fast of service they're giving to their end user what's over
25:17
subscribed and not over we won't talk about dirty secrets
25:20
>> yeah well keeping keeping our network um you know running at the right I guess,
25:27
you know, at yellow versus running at red is is is a challenge. Um, you know,
25:32
I'm not gonna I'm not probably skilled enough to to technically to say, okay,
25:36
this is how I see the peering changing. That's not, you know, my strong suit.
25:41
But, you know, when I look at the the network map and and we see this stretch
25:46
on yellow, this stretch on red, you know, it it really comes down to uh
25:51
proactive management, understanding where your traffic is coming from, what
25:54
you've sold, and then managing those vendor relationships, and whether it's a
25:57
fiber provider, whether it's our own network we have to overpull, whether
26:01
it's, you know, a situation adding, expanding a a DWM, you know, uh system,
26:07
it's all about proper planning and and owning your own global backbone for 30
26:13
years and being all you've done. You get really good at that anticipation and the
26:17
planning and it comes to when we get certain deals that are in you know
26:21
Salesforce. I as a sales guy I live in Salesforce right
26:24
>> so when it gets to stage four that's when my planners start to to be aware of
26:29
it and if you know it runs different scenarios what if we win this do we have
26:33
the capacity yes no should we start planning for this now so it it can be
26:37
anything from um you know the the CDN networks it could be broadcasters it
26:43
could be you name it so again I I can't get into the peering conversation too
26:48
deep but what I can tell you is managing uh available capacity and making sure
26:53
that we always have enough overhead
26:54
>> to to accommodate all the traffic that we've sold and all the burst traffic we
26:58
could get. Um, I'd love to tell you it's a real easy thing to do, but it's it's
27:03
an art as much as it is a science. And and I leave that up to the smart people
27:06
that have been running our network for 30 plus years. there is a um there's a
27:11
personality like an old old salty sales personality where it's you know they'll
27:15
talk about like the race to zero with telek
27:18
>> and and networking bandwidth and and on one hand I appreciate the sentiment
27:22
because it used to be you could make you know like T1s were very expensive once
27:25
upon a time I sold a T1 it was egregiously expensive or installed a T1
27:31
and you know and if you're selling a gig circuit gig circuits once upon a time
27:35
were expensive and now are you know but you're looking at like well okay you're
27:38
selling a a you know gig e circuit and we're selling 800 gig circuits and like
27:43
there's a big different you know delta between those two. um how you know that
27:48
that creates a lot of pressure also you know for you and your sales organization
27:52
right because you've got to create new capacity people are looking for more
27:56
bandwidth it's like does it ever feel like you're off the wheel of like cyclic
28:00
upgrades and you know ever increasing demand and capacity you're not I mean if
28:05
if you talk to anybody you know you look at a lot of the competitors in the
28:09
market it it's a you know the capex budget is always got to be there and
28:13
you've always got to be looking should I bill should I high. Um, and then all the
28:17
equipment that's on the ends of those those stretches, the fiber. I mean, it
28:21
is a wheel. And I'd love like again, I'd love to say that it's a you know, you
28:24
get to a point it's just okay, but it's you're you're always you're always
28:29
trying to figure out where should I put my money? What's the best investment I
28:33
can make and return on that? Be it a new fiber stretch, do I the same route? Do I
28:39
look at a new route? Because the things that the big companies, the
28:43
hyperscalers, the the the the cloud companies, it's about less about the
28:49
price. It's more about the speed to deliver and the diversity of the path.
28:54
All that means capex,
28:55
>> right? So it is absolutely you said it right. It is absolutely a wheel that you
29:01
have to manage on. Okay, if I put this here, it can't be if you build it, they
29:06
will come. It's got to be I know that there's companies that have a demand for
29:09
this. we've seeded the market with this. We're going to we're going to make it um
29:14
and and and make sure that we can not only maintain what we have, but also be
29:18
interesting to those next wave of of users coming on because everybody wants
29:22
different markets. I mean, there's you you look at there's a lot happening in
29:26
the Nordics right now. There's a lot happening in Iceland right now. There's
29:29
I mean, it's it's five years ago if you'd said those are the hot spots for,
29:33
you know, new data center builds and fiber, no way. But it's all about
29:37
chasing the power. I mean there's a lot of dynamics to the market but um it it
29:42
is a constant conversation that we have about okay here's the capex budget we
29:46
have to keep the lights on turn this this uh red to yellow or green this
29:51
stretch so there's there's maintenance of existing traffic but also how do we
29:55
continue to grow the network and be attractive to you know that next wave of
29:59
customers who are saying I'm unhappy with this legacy provider I want to look
30:04
at an Aurelion what do they have that's unique and better and why it's I'd love
30:09
to say it's it's it's a it's always there. That's an interesting dynamic to
30:12
talk about. the original I don't want to say original
30:17
but for a long time um network builds followed major metro markets right so
30:23
you had the you have center gravity of the internet uh west coast the bay area
30:27
east coast you know ashurn area
30:29
>> of course
30:30
>> but then everything else was whether it was related to a cable landing station
30:34
you know system right so Miami versus Seattle you know west coast you know New
30:40
Jersey right so you had where the cables the Oceanana cables are coming in or you
30:44
had like you know Atlanta is a huge market so you end up with data you know
30:48
infrastructure in Atlanta then you have you know like oh it's cheap land we're
30:51
going to build data centers there and we just have so there's there's this but
30:54
now you know you talk about data centers you know if you're doing a gigawatt data
30:58
center project this isn't going into like downtown New York it's it's going
31:02
into places
31:04
>> that haven't had any built before
31:05
>> Grand Prairie Texas
31:07
>> right exactly you know or or you know you know Oakliff South you know Red Oak
31:11
or you know I mean farther. I mean like like middle of nowhere projects, right?
31:15
So data center companies have to compete with you know cost of capital, location,
31:19
entitlement, stability of local government and and um you know
31:24
geopolitics, right? Can they get the power? How much does a power cost? But
31:28
so you get all this like stuff to entitle and then all of a sudden you
31:31
know you turn around you're like oh we're going to put in this you know
31:34
gigawatt data center project that's not going to be like a 100 meg circuit going
31:37
into it. that that throws your capacity planning for I don't want to say a loop,
31:41
but like all of a sudden now you're looking at like, okay, there's going to
31:44
be a lot of something happening in this facility. Um, how far ahead are you
31:49
engaged in those conversations and how much of that is, you know, does this
31:53
become a strategic customer, strategic site for us? Are we trying to stub this
31:57
off of something else? You know, are we going to see a lot of these like stubs
32:00
coming off of cities versus like, you know, primary network nodes? Yeah, I
32:04
mean usually when we're involved in these things, it's it's driven by a
32:09
large customer build, right? And a specific purpose-built data center. And
32:14
you know long term you have to decide is this is this something we want to do
32:19
because generally we've made our money where we say we're in this this area
32:23
this pop this data center pick your your terminology because we know that there's
32:28
30 different customers there and there's another 75 100 a thousand that we don't
32:32
have that we want to win out of that. Right? when you when you do a
32:36
purpose-built data center, you got to make sure that it really makes a ton of
32:40
sense because the money it costs to build into that all the things you
32:44
mentioned is is not small and it's got to be a long-term contract and okay, am
32:50
I able to also have other customers leverage this so I get the best bang for
32:55
my buck. It's, you know, we don't do a ton of those. We do a few of them and uh
33:00
we we take those um those decisions, you know, uh very seriously because it's a
33:05
huge commitment. Um but I think you're going to see more
33:10
spurs off of networks or you're going to see um I was trying to think there there
33:14
was a deal we did in North Dakota uh for a large customer and um we had to
33:20
basically resell a local provider who happened to also be a customer. So, it
33:24
was a lot. That's one of the values we bring. And we bought we bought network
33:28
and they had to do a small build to get to our closest pop, which was not in
33:32
North Dakota. But, but it but then once it hit our backbone, it it was great for
33:38
us. So, we did a you know, you think, well, that sounds like an enterprise
33:42
deal where you buy a a tail and you know, but it was over bigger
33:45
geographies. We're seeing more and more of that from the big hyperscalers to
33:49
where they want to be in a certain geography for whatever reason. Power
33:52
space, pick pick it. um but there's nobody there so they they have to call
33:57
the local provider and they don't want to mess with that. But we know those
34:00
local providers, they're all customers. So we we we tie those guys together to
34:04
get to our closest pop. So we're seeing a lot more of that and I think that's
34:08
going to be continued to be the the the norm versus everybody wants to be in 60
34:13
Huts or in you know the Infomart or in
34:17
>> what is it? Wilshshire. Yeah, one Wilshire. So I I think we're going to
34:21
see more and more of that. um kind of spur or kind of off the beaten path back
34:27
into the main artery of the of the uh the network that's existing.
34:31
>> How much do you think um data sovereignty is driving conversations and
34:35
selections now of locations for customers?
34:37
>> I mean, it's always a part of the conversation and everybody's got, you
34:41
know, it's like anything else. People have uh very strict people have a little
34:46
more loose. So, it but it's, you know, it's not something that I deal with. I I
34:49
I make sure that our our uh our attorneys and our folks deal with that
34:53
and make sure we we cover our bases, but it's always part of the conversation
34:56
these days, as is ISO certification, as is, you know, all the security things
35:01
that that are happening in the industry today.
35:03
>> Yeah, data sovereignty seems like it's going to start ratcheting up big time.
35:06
It feels like there's,
35:07
>> you know, certain geopolitical alignments changing and shifting and,
35:11
you know, pressure to have things in certain markets.
35:13
>> It's never dull for sure.
35:15
>> Yeah. Layer eight issues are always fun to deal with. It's like I just want to
35:18
build a network. Do I really have to deal with this?
35:20
>> Two and three, maybe one.
35:21
>> Yeah. Yeah. No, I just I just want to I just want to make the lights blink, you
35:24
know? Can I just make the lights blink, please?
35:27
>> This show exists because of what we do at itroker.com. If you're in the middle
35:31
of a real tech decision right now, new technology, vendor selection, a contract
35:35
that doesn't feel right, an M&A event that just landed on your lap, and so on.
35:39
We help buyers like you get it right. Independent strategy, sourcing, and
35:42
contract negotiation. No kickbacks, no sales quotas, just someone in your
35:46
corner. schedule a call at itbroker.com. Back to the episode.
35:52
>> He said something a little bit ago that I want to venture back into. Enterprise
35:56
customers have um different demands and pressures of of how they acquire
36:01
technology. And you know, there was a promise for cloud
36:06
for a long time that that um I don't even know how this seated out. It was
36:10
like it's going to be cheaper for you to be on a cloud. And we all know that
36:12
cloud is not cheaper. Like anybody who's ever done anything with cloud,
36:15
>> nobody will tell you they're buying cloud because it's cheaper. They're
36:17
getting other things out of it. Flexibility
36:20
>> scale. They're displacing their entire procurement load onto something else
36:24
that nobody can actually deal with. It's like, oh yeah, you know, we're spending
36:27
this much money here. And nobody's like, "Okay, spend less." They're like, "We
36:29
can't. Sorry, we have to deal with you anymore." Um, but then now a lot of
36:34
these projects where we're talking about enterprises getting back into control
36:37
and taking control of their infrastructure again.
36:39
>> Yeah. You know it. Some of that conversation starts with cost. Some of
36:44
that conversation starts with performance. Some of that conversation
36:47
just comes from like budgeting and predictability. And that third one has
36:50
been very interesting for me to watch as well of like how much of the
36:53
predictability is actually driving a lot of like you know cost versus
36:57
performance. Um when you're when you're when you're engaging with a with an
37:04
enterprise customer and you know 400 gig circuits we're talking you know this is
37:07
significant network build. How much of it is performance? How much of its cost?
37:11
How much of it now is just saying, "Hey, we know we have this capacity. It costs
37:15
us this much money. We can plan around it. We know what we're working with, you
37:18
know, and and getting back into this idea of like, you know, owning and
37:22
operating our own network again."
37:23
>> Yeah. So, I think the the first the first kind of lever on that conversation
37:28
is the cost, right? To your point, um, they've put all this stuff in the cloud.
37:33
Yes, they can scale, but the predictability is is is is hard, right?
37:38
And it's it's a big number. And so then they say, what's the what's the best way
37:42
to utilize this? It's kind of like the wireless conversation. There's
37:45
definitely a use case for it, and there's obviously a use case for cloud,
37:48
but you think about, okay, what do we need to put in the cloud versus what do
37:51
we need to take back or put someplace else? And I think those are the
37:55
conversations that we see more companies having because it starts with cost, but
37:59
then there's a, you know, I still need to be able to scale. um I still need to
38:03
be able to do things quickly. We say, you know, maybe put your your your
38:06
storage and some compute in the cloud and put other things applications in in
38:11
different data centers. I leave a lot of those conversations to the engineers,
38:15
but I think to your point, the the cost of cloud and the extreme, you know, it
38:21
can spike and is what's making people step back, enterprises step back and
38:25
say, is this the right use of this this spend and are we doing this correctly
38:32
and maybe maybe not and so I think those are the conversations we're involved in
38:36
where they say we want to move this from the cloud to our, you know, big HQ site
38:41
or to this data center site and that's going to change the way our network
38:45
typology is is is working. Can you help us with this? more like absolutely but I
38:50
bring in the smart people for those conversations
38:54
>> getting into cost conversations with cloud or performance conversation with
38:57
the cloud people get very focused on cost of compute you know cost of storage
39:03
>> and for me the relative cost for compute or
39:07
the relative cost for for storage I mean you can you can build your own
39:10
infrastructure have bare metal you know there different tiers of it and you can
39:13
get a more efficient compute cost right but it's
39:16
>> that's not the big lever for me And and I'm, you know, trying to explain
39:22
somebody the difference of purchasing 100 gig on your network versus paying
39:27
for 100 gig on a cloud provider like the delta is so extreme that people
39:34
don't believe it's actually true when we get into this psych and this
39:38
conversations. So how are you managing that cycle? Because I find it very like
39:43
it's hard to be upfront and be like, "Hey, by the way, this is what you're
39:45
paying and this is what it would cost to do an alternative." And and then, you
39:48
know, I get I certain times I get responses, they're like, "Oh, you're
39:52
lying to me." And I'm like, "No, I promise this is actually here's
39:54
paperwork. I can prove it." You know,
39:55
>> but the cloud's so much cheaper.
39:57
>> Oh, no, no, no. I mean, you know, I mean, look, everything's negotiable,
40:02
but like your base rate on egress and cloud is 5 cents a gig, right? So, you
40:06
know, the cost differential from that to an IP transit link,
40:10
you know, 300 times more expensive, you know, like it's
40:15
significant, right?
40:16
>> Yeah. I I I I personally don't get into those conversations when it at that
40:21
level. Usually, companies have made a decision one way or another and that's
40:25
when they bring us in or that's when I've been involved. So, I I can't really
40:29
speak to that to say, well, this is why somebody would do it. I mean, part of
40:32
part of the conversation is, okay, your your your cloud provider is is uh is not
40:38
a network provider. You know, they've they've got a cloud fabric that they've
40:42
stitched together and they can do a lot of things, but then where does that
40:45
traffic need to go? How does it traverse the internet? That's really where our
40:48
strength comes in. But from a cost perspective, no argument there. I just I
40:53
you know usually we we're not faced with this putting a hundred you know au 10
41:00
gig access in the cloud versus buying 10 gig of IP transit tell me why I should
41:06
do this those conversations are well pasted by the time we get that they know
41:09
what they want typically the interfaces on these clouds you know for direct
41:14
connect or express routes I I mean they're way behind industry performance
41:19
in terms of Ethernet speeds and I don't know how much of that is just
41:25
I I don't know how much of it is uh intentional to you know keep people you
41:29
know captive to a certain degree and how much of it's just that just they're so
41:32
far behind getting these network links up um you know given that like all these
41:39
cloud companies are your customers for your IP transit and you know and
41:43
transport networks as well as you're trying to deliver service for your
41:47
enterprise customers to connect to them you know do you see this friction
41:51
continuing doing or you know should we continue to expect you think for people
41:56
you know for the cloud networks to be you know you're rolling out 800 gig I
42:00
you know how long before we see 400 gig available you know for private
42:04
connections
42:05
>> yeah great question I I can't you know we we provide a cloud connect service
42:09
and that's as far as we go into the cloud so so I I can't say what's on
42:14
their road map but um but yeah the price erosion is and and price compression is
42:20
always there and it comes from different places uh not only competitors that are
42:25
that are uh com you know that that are in our same space but also from you know
42:29
the cloud providers looking to sell as well. So but I I can't speak to that as
42:34
much as I probably if I have my engineer here maybe I could I could do a bit
42:38
better.
42:38
>> Yeah well
42:39
>> I told you I'm not an engineer I'm I'm trying I'm trying to I'm trying to walk
42:43
the line here carefully. um network as a service is becoming a term, you know,
42:48
that's it feels like it's taken hold now in terms of like a segment and you know
42:53
and I've seen some companies selling network as a service where it really
42:56
feels like they're trying to create a new product category that they can have
43:00
different commercials on and not necessarily actually delivering like
43:03
service value to the customer. Um, and I see other network as service companies
43:08
coming online that it's that are that are really trying to create this like
43:10
um, you know, again back to like that SDWAN
43:14
magic. It like feels like an iteration of SDWAN.
43:16
>> Yeah.
43:17
>> Where we can do all this stuff for you and you don't know have you don't have
43:19
to know how it works. when when you're looking at and people
43:24
are, you know, looking at this and you're you're saying, "Okay, well, in
43:28
most cases, it's probably writing on your network, you know, and there's some
43:31
overlay on top of it, you know, is this something that you now have to chase and
43:36
figure out like what's the actual economic driver or or you know, customer
43:40
requirement that's driving like why is NAS all of a sudden becoming a thing?"
43:44
>> Yeah.
43:45
>> And you know, you some of your competitors, I believe it's a thing
43:49
because their contracting cycles are so difficult to deal with, right? like I
43:53
can't get I can't contract with you easily. I can't install with you easily.
43:56
I can you know like it's not easy to work with you so therefore I have to go
44:00
do something else.
44:01
>> Yeah.
44:01
>> Yeah. I I think that I think that everybody's looking for their value
44:05
proposition in the space right and I think NAS network as a service. I think
44:09
a lot of it the core of it is the ability to scale quickly.
44:12
>> Um yes there's the analytics piece. Um, you know, I I I think that
44:20
I I think that it it really depends on on what the enterprises wanting to do
44:24
themselves and what they want to contract out for. Um, you you pay you
44:28
pay premium for all those services
44:30
>> and and so you see, you know, I like I'll liken it to an SDW sort of uh life
44:35
cycle that we've seen. Most most companies started out buying uh SDWAN
44:40
from a managed services provider. So they would buy, they would, you know,
44:44
the company would install, they would do the underlay, they do the overlay, and
44:47
you'd pay a premium for them to manage it. And over three-year term, sixyear
44:52
term, the IT teams have kind of gotten wise to this, and they've said, you
44:56
know, we can we can manage these things and we may or may not need uh managed
45:01
services provider because we can maybe roll our own. So on the second or third
45:05
iteration of SDWAN, they've bought their own boxes, they've bought their own
45:08
software, they're managing it themselves.
45:12
um and they've made that choice over time. Some companies are absolutely
45:15
still with managed services providers for the same reason. They don't want to
45:18
manage it, but you pay a premium for that. So for us, it's really about um
45:22
qualifying the opportunity to say, look, we are an underlay provider, and that's
45:26
what we do. And we can scale really quickly. Our processes are this. If you
45:30
want all the other bells and whistles on top of that, you can either pay for
45:34
SDWAN yourself, buy somebody else. We can play with the, you know, we can play
45:38
work well with others. But this is what we do well. And so network as a service
45:42
providers have a different proposition than I do. Um we're pretty
45:46
straightforward with the connectivity. Our ability to scale and deliver fast is
45:50
is is one of the reasons we've been successful globally. Um and so to me
45:55
when I hear network as a service, yes there's the monitoring and some of the
45:58
other bells and whistles, but at the underlying it's the ability to scale and
46:02
make it easy for the customer. Um you know we do that anyway. So if they want
46:07
more then then it's really about us qualifying the opportunity saying what
46:10
do you really want like we will lose a we will lose an opportunity because some
46:14
customers will quote I will do this with a NAS provider or an SDWAN provider who
46:20
does managed services and then I'll do another business case that says I'm
46:23
going to roll my own devices my own software on SDWAN or Sassy or whatever
46:28
but I'll do the underlay separately
46:30
>> and and some we win some we lose. It just depends on how they want to spend
46:32
their money. I don't I don't know if that answers your question, but it's
46:35
>> I get it. Uh common in the network space, backbone carrier space is
46:44
um growth by acquisition, you know, so we see, you know, and this has been true
46:50
for a long time, right? So growth by acquisition or um buying customer
46:55
traffic.
46:55
>> Yep.
46:56
>> Through price points. Um aggregators notoriously do this, right? where they
47:02
have a they have to sell to a certain point or they have cap committed
47:05
capacity and whatever they have undersold like it's either take the loss
47:09
on it or sell it for anything you can right creates downward pressure on the
47:12
market and um those companies end up in a life cycle that either ends with
47:18
bankruptcy and a reorg which also has devestature of assets or
47:23
like you know there's some big networks right now that have way too high of a
47:28
debt load and are trying to avoid the bankruptcy and are shedding assets as
47:31
quickly as
47:32
>> yes
47:32
>> they can find anyone to buy them. How do you stay focused on your core business,
47:38
you know, and not get wrapped up into the craziness? I mean, you know, Telly
47:43
Aan has been doing this for a long time and you've managed to stay out of this
47:47
craziness. How much of that is just, you know, corporate lineage and and culture
47:51
and how much of that is just like every day, you know, like, oh, this company's
47:54
doing this thing. Do we have to chase them or not? Or, you know, like,
47:57
>> well, we've tried before, right? it it hasn't been a big big scene but we have
48:01
quietly tried to do some other things to diversify that revenue stream to offer
48:05
more of the stack but at the end of the day I think you know prior to me joining
48:08
the company I've been here four being at a railing for four years now I mean the
48:12
decision was look we're going to take place our bet that connectivity is going
48:17
to be the underlying most important thing in the stack
48:22
>> and we're not going to play in the race you know the voice game the security
48:27
game the SD and you know, we're just not going to do that. So, it was more of a
48:31
corporate lineage to say, let's stick with what we're good at, uh, and not try
48:35
to be all things to all people. And, you know, knock on wood, so far that's paid
48:39
off with us. Um, one of the big value propositions that we have compared to
48:44
the market, like I mentioned before, is the network performance. When you
48:48
acquire multiple networks, you acquire networks that were never designed to
48:52
work together. And you've got all kinds of technical reasons that why could be
48:57
vendor problems, it could be um different ways of doing signaling across
49:02
disparate networks, but they were never designed to work together. So when you
49:05
place that across a global need, you got performance issues, you've got latency
49:09
issues. I mean, you've got all kinds of issues.
49:12
>> We have stayed true to what we do. So every expansion we've done was designed
49:17
just like the other components of our network. So it all works together well.
49:20
So over time I've been able to separate myself to say I'm the best global
49:24
provider there is in terms of delivery experience. I'm the best global provider
49:28
there is in terms of network performance because my network is you know is is uh
49:34
homegrown and it's ubiquitous and it's designed to work together. I've never
49:38
tried to be a voice, an SDWAN, a security, a managed device, you know,
49:43
company. I've stayed true to what I what I'm I'm good at and and we are the best
49:48
at that. So, it's kind of been a lineage choice, but I think long runs it's
49:54
seeming to pay off because like I mentioned, a lot of these things are
49:58
getting commoditized, voice, SDWAN, all these other things. whereas the only
50:02
thing that's really growing in the stack is the need for the network and AI is
50:06
only driving that. Cloud has been a huge driver for us as well. So, so um yeah,
50:11
there's we have to qualify part of that, right? Because you say we're not a
50:14
security company. You're not a security company in the sense that you're not
50:17
providing firewalls for your customers. You're hanging off their networks. But,
50:21
you know, backbones have their own security issues, right?
50:24
>> We we do provide DOS. Yes. So, so we do that.
50:27
>> Let's let's talk about DOS for a little bit, right? you know cuz I can again
50:31
been doing this for a little bit right and you know I remember the days where
50:34
it was like oh it's a 10 gig DOS a volumetric attack at 10 gig and you're
50:38
like oh wow that's so big what are we going to do about it right and and you
50:42
know and so this is this is one of those like uh double-edged swords for me it's
50:45
like oh your backbone just went from a 10 gig to 100 gig to 400 gig 800 gig
50:49
we're going to be over a terabyte
50:51
>> in a heartbeat right like
50:53
>> and and and you look at that and you're like okay great so the unintended
50:56
consequence of course for this or like eyeballs Right? It's like, oh, you know,
51:00
we're we're upgrading, you know, to giggy 2 and a half gig, 5 gig, you know,
51:03
10 gig pretty soon to the homes, right? Like that has unintended consequences as
51:08
well that comes down the line. DOS is a very crazy animal to deal with
51:14
at your scale. Um, so, you know, I'm I'm going to leave that as like a a leadin,
51:20
not actually, you know, get into specifics just yet. I'm curious your
51:23
response to this is you know again somebody you know who has to provide the
51:27
capability and protect the customer and then sell this as a service which at one
51:31
point became like I think DOS was here's actually the first part of the question
51:35
DOS was like an optional thing that people were were not buying and I look
51:38
at it as like very much a like this is core capability like you know now
51:42
>> what's saying security is step zero in building a network right
51:46
>> so our DOS solution is something that we built and use internally
51:51
>> and so we've decided to sell it as well uh on top of you know services um five
51:57
six years ago. So it's absolutely something that is native to our core. So
52:02
the solution that we sell is the same solution that we provide and we can
52:07
recognize an attack and and mitigate that attack very quickly. And you know I
52:13
can't get into specifics but you know 10 gig is nothing. Yeah,
52:18
>> nothing that and and so what we've seen is some of the biggest in the world and
52:22
we've been able to quietly mitigate those for our customers that it's
52:26
happened to. So, you know, you never want to talk too specifics about
52:30
security and all that stuff about customers. But I think what I'm curious
52:34
about is
52:36
>> when we're selling and helping deploy large links and IP transit links, part
52:41
of the conversation very quickly turns into what's your DOS strategy across
52:45
this link, right? and um and I and I feel like it's it's
52:50
finally gotten to the point where people accept it like we we understand that we
52:53
need to do something to protect it here. How I don't want to say like what
52:59
percentage of like buying DOS versus not but like how much of that conversation
53:03
shifted for you over the course of the last like four or five year you know
53:06
three four years of like you know is that still a lot of tension in the
53:09
process or is it like hey do you have DOS do you want ours do you have
53:11
somebody else's like what are your what's
53:13
>> that's exactly it I mean what what we we always recommend it right as part of the
53:18
package and uh sometimes we include it depending on you know how much the
53:22
customer is spending and their need for it but a lot of customers already have
53:26
it. and they have it as part of a different part of their existing stack
53:31
and we say get it but the way that ours works is you only pay when we you know
53:36
there's a small monthly fee but you only really pay a meaningful amount when
53:40
we're mitigating an attack. So it's more like a very good insurance policy that
53:45
you only need when it happens whereas other companies have different business
53:50
models. So we always we sell it a lot more than we used to because it's more
53:55
prevalent. But a lot of companies already have that box checked and they
53:58
don't want to talk about it because it's part of a broader security strategy.
54:01
>> Yeah. And also within that, right, you talk about like um charging to mitigate
54:05
an attack, you're really talking about what what what's going through a
54:08
scrubbing center, right? Because if you can block an attack on the cloud or
54:11
whatever.
54:11
>> Yeah. If appear if you can block with a rule on appearing, you know, on an edge,
54:15
right? Like that's, you know, a lot of attacks are like, "Oh, we can just block
54:18
it on our edge and boom, you know, push out, you know, flowspec rule and and
54:22
you're done, right?" I mean, I make it sound really simple, but of course.
54:25
>> No, but but to your point, I mean, there's companies that specialize in
54:28
that and they market heavily to that. And so, it's part of our fiber as a
54:32
company and and it's not something we've said we're this big security company.
54:36
We're a connectivity provider, but it is part of our ecosystem. And so we're
54:40
we're selling it more and more, but it wouldn't be the thing that
54:44
>> I put as my front, you know,
54:46
>> we're known as the largest IP transit company in the world, right? You know,
54:51
single, I think, top four wave provider in the US, you know, all these things.
54:56
AS1299 is the highest quality more so than, hey, we have this great DOS
55:00
product, right?
55:00
>> No, of course. I mean, that's it. It it used to be like DOS was the product
55:03
where people were out buying it and it it I'm I'm actually really happy to say
55:07
this. I'm really happy to say that it's it's moved into it's like you have a
55:10
lock on your house, you have an internet connection, you have DOS like it it's
55:13
>> it we it feels like we finally hit that point because there was just a decade
55:17
plus of pain, right?
55:18
>> Of course.
55:19
>> Um and I and I'll say like working with, you know, I I'll use this term like a
55:25
security engineer, you know, that's going through and doing and you know and
55:30
filtering DOS attacks are some of the most enjoyable conversations I've ever
55:34
had, right? right? In terms of like this the quality and the skill set of the
55:36
people you're interacting with now
55:38
>> in the time you know the the joke for us you know network engineers it's like
55:42
denial of sleep attack right like it's not you know like when we go back and
55:46
like I'm I I talk to people that want to deploy some sort of like on premise of
55:49
their own DOS system and I'm just like you're out of your you know like I have
55:52
to be polite but like I'm in the in my mind I'm thinking you're out of your
55:54
mind like you have like this is nuts don't do this
55:57
>> right
55:58
>> um you know the other part of it that I don't think people really understand or
56:01
appreciate is you have a lot of customers.
56:06
So your obligation is providing you know a performant
56:10
product for all of your customers and there is a point where if a customer is
56:15
degrading that experience you know the blast radius is affecting
56:19
other customers you know somebody has to make a decision be like you know
56:22
>> absolutely you know I mean those are all covered in MSAs and things like that
56:26
because at some point you you have to do what you need to do to like you said
56:31
protect your other customers but you know it's something that we monitor and
56:34
use internally all the time, 24/7. It's part of, you know, when you you you
56:38
can't do all the security certifications without proving that you can do all
56:42
these things. So, um, but like I said before, to say that people know us from
56:47
our DOS service, no, but it's absolutely part of the conversation. I think we the
56:52
uptick when we sell network is is is improved. We actually bolted it on to
56:57
it. We have a like uh um it's called IP secure where it's a it's a package deal.
57:02
We sell a lot of that. But to say it's uh you know 70 80% of our sales, no. But
57:09
it it is an extremely good service. And uh
57:12
>> the reason we don't sell it more is because a lot of companies are already
57:16
locked into an agreement with a long-term contract. It's part of a
57:19
bigger uh you know uh strategy. So
57:22
>> it's like great. You have it. Fantastic.
57:23
>> Yeah. Exactly.
57:24
>> You know, like little PSA commercial here is you know, if they're not on your
57:27
network and they want to come by your DOS, it's not like something that
57:29
they're going to go and provision in 10 minutes. But if they're on your network
57:32
and they're having an attack and they haven't provisioned DOS with anybody,
57:36
they can get yours pretty quickly. Like this isn't like a we're going to spend
57:40
days and weeks and months deploying something. It's like, oh, we have an
57:43
issue. We need to resolve this issue. And your knock team can be like, okay,
57:46
yeah, we can facilitate this pretty quickly for you.
57:48
>> That's right. Once once it's up, it's very easy to use. And it's not and like
57:53
anything, it's not fire and forget, but but it's something that we we uh we do a
57:57
very good job at. and we have thwarted some of the the biggest attacks that
58:00
there have been across the internet. Again, I can't say names, but big stuff.
58:04
So,
58:05
>> you know, I don't want to um I don't say this to like,
58:12
you know, network engineering, backbone
58:15
engineering, network capacity, right? I don't want to say like it's a solved
58:19
problem. I'm not I don't want to position it like that because you know
58:22
going from 10 gig to 100 gig introduces like all sorts of new problems across
58:26
the entire environment. You know going from 100 gig to 400 gig, 400 gig to 800
58:30
gig. Now you know after 800 gig I mean this next decade of growth
58:36
you know the capacity curve is going to be just absolutely staggering to watch.
58:40
>> Yeah. the
58:45
you know and then like the demands of enterprises coming and coming back and
58:48
becoming a core customer of yours like that changes this game a little bit for
58:52
you as well. So I mean as you're looking forward over the course of like the next
58:55
year next two years I'm not I don't want to phrase it like what challenges do you
58:59
see or what do you have overcoming but like what's been surprising in that like
59:02
planning process of thinking through and and trying to get ahead of a little bit?
59:06
Well, it we we've had to to be much more proactive with having uh you know,
59:13
equipment uh caches available for cards and expansion very quickly. You know, to
59:19
your to your point, um you ask some of the social networks or hyperscalers for
59:26
a forecast,
59:26
>> capacity planning,
59:27
>> they they say pick a number,
59:29
>> you know, I mean, honestly, so so you can only do what you can do and and you
59:33
have to monitor it. And it's something we've gotten very good at in terms of
59:36
understanding uh what's coming. We communicate with our customers all the
59:40
time and we see uh big spikes or big drop offs. You know, it's it's really
59:45
just staying on top of it. But to say that we have the magic formula to, you
59:50
know, problem solve, no. Because, you know, five, six years ago, we weren't
59:54
even talking about AI workloads and that's changed the game completely. Um
59:59
what it's done for us is we've had this big massive global network for years and
1:00:03
years and years and uh we're looking how do we we always say we need to fill up
1:00:07
the suit you know how do we fill the pipes if you will
1:00:11
>> um and this is really I mean this is played in our favor in a lot of ways but
1:00:16
it's not stopping.
1:00:18
>> So you know like ideally you would run your capacity at like 100% and there
1:00:22
would never be a spike that would go over 100%. But then that's not the way
1:00:25
that the internet works. It's like you have to run way lower capacity and and
1:00:29
and and you you touched on supply chain briefly. You know, CO changed the supply
1:00:33
chain game for a lot of people, right?
1:00:34
>> Well, there's been many things that have impacted the supply chain.
1:00:36
>> Yeah. I mean, memory prices right now are changing the supply chain, right?
1:00:40
So, but that's also an interesting challenge for you, right? Because you're
1:00:43
not talking about like trying to figure out how to do parts depot in one
1:00:46
country. No,
1:00:47
>> now you're talking about like how do you do on a continent by continent basis do
1:00:50
parts depot.
1:00:51
>> Well, and that's it. I mean, we have a strategy there to where globally we can
1:00:54
get anything there in a matter of days. in some places ours. But but the the the
1:00:58
the real dynamic that's that we've seen is that
1:01:02
>> you know a lot of the larger hyperscalers and companies like this are
1:01:06
going directly to these manufacturers. And so we're not competing with our our
1:01:10
you know other carriers for equipment. We're competing with right
1:01:14
>> hypers and so they got deep pockets and big checks. And if I'm I'm an OEM an
1:01:19
equipment manufacturer, well that going where that money is. So you throw any
1:01:24
any disruption in the supply chain uh distribution manufacturing pick a topic
1:01:30
we all get hurt. So, so you just again for us it's a matter of we're in a good
1:01:36
spot now because we've had this massive network for years and we're starting to
1:01:40
really get a big use out of it. But now we have to continue to grow and we have
1:01:45
to continue to expand in those spurs like I mentioned in other areas overpull
1:01:49
in some areas that are now hot hot places for data center stuff. And so um
1:01:55
you're trying to get fiber, you're trying to get equipment, you're trying
1:01:57
to get construction permits and all those things that happen. It's not easy
1:02:01
layer on top of the supply chain issue, you know. So, you you just have to stay
1:02:05
as as ahead as you can. You have to forecast accurate and somewhat
1:02:10
conservative delivery timelines to your to your customers when there's
1:02:14
construction or there's equipment issues and uh you know, you have to be
1:02:17
transparent on what the situation is.
1:02:20
>> So, that's a it's an interesting horse cart conversation, right? So
1:02:25
hyperscalers come in, big enterprises come in and they have a very simplistic
1:02:32
requirement set. You know, they want very fast, low latency Ethernet
1:02:35
connectivity. Right now you get into, you know, how are you actually building
1:02:38
out your core leaf at spline? You know, are you doing BGP based, you know, uh,
1:02:43
VPN architectures inside of the data center in order to deal with containers?
1:02:45
Like there's there's there's a there's a whole different set there, but they
1:02:48
don't need MLS. they don't need EVP you know EVPNs they don't like it's a
1:02:52
simplistic set right but at the same time because of the um
1:02:59
uh you know the the inventory the parts that you know how much what they're
1:03:03
ordering right that's good because the chip manufacturers now have a pipeline
1:03:07
of manufacturing which drives down prices on those chips that become
1:03:10
available right so like there is a certain curve where you're like okay
1:03:12
demand's crazy but cost per chip can now decrease because volume capacity
1:03:16
increases but then you've got the OEMs right that actually build
1:03:21
>> right
1:03:21
>> infrastructure that together
1:03:23
>> there's not very many of them
1:03:25
>> correct
1:03:25
>> there's consolidation in the market taking place right which we'll still you
1:03:29
know be interesting to see how this plays out and so then you've got another
1:03:33
like you know horse and cart issue where you need to grow your capacity there's
1:03:37
not very many OM that can build that capacity but like how much of that
1:03:41
capacity they're building for the network providers I mean you're buying
1:03:44
lots of gear
1:03:45
>> lots and lots and lots of gear right so you're a dominant customer for them and
1:03:48
there's a few other networks that are buying lots and lots and lots of gear,
1:03:50
but like compared to the enterprise customer, you know, they buy lots of
1:03:54
gear as well. And so like their product plans are very, you know, like that's a
1:03:58
fascinating, you know, problem to deal with as well. It's like how much how
1:04:01
much of this do you think at this point is like the horse driving the car, you
1:04:04
know, the cart or the cart's driving the horse or, you know, like what's the push
1:04:08
and pull like?
1:04:09
>> I mean, that's a tough spot for them to be in right now, right? Because, you
1:04:13
know, I talked to some of our equipment um manufacturer guys and they say every
1:04:19
morning I wake up and there's four escalations because somebody, us
1:04:23
included, needs more stock and they have to figure out how they make all that
1:04:28
work. Um and that's been the last year. So it's a fair it's a rarely I mean
1:04:34
supply chain issues during COVID aside
1:04:37
>> um the dynamic of who's buying and the volumes they're buying nobody can you
1:04:42
know just like they can't make a network you know a traffic forecast they can't
1:04:46
make an equipment forecast um you know a lot of the the the bigger the the good
1:04:50
only good news is you know the data center builds take a long time right
1:04:55
you're not slapping those up in 90 days those are multiple years y
1:04:58
>> right and so you can kind of plan for those it's it's the um it's the ones
1:05:04
where they want to buy you know 4x400 gigs and they need it installed in 90
1:05:10
days and 60 days and you're like oh my you know it's usually it's 60 days and
1:05:14
we have enough stock to get here but we need more you know it's those things
1:05:18
that challenge and it's not just us it's all that carries so I I don't have a
1:05:22
great answer for you I know that we put a lot of pressure on those guys and we
1:05:27
communicate with those guys on a weekly basis because the forecast changes the
1:05:30
needs change and um yeah, they're in a tough spot. I mean, it's a good problem
1:05:35
to have. We just have to produce more so we can sell more, but there's challenges
1:05:39
with that, too. But I I don't know the right answer and and uh you know, how
1:05:43
that's going to play out. It's, you know, you can't have one vendor, you
1:05:47
can't have two, you got to have at least two or three for this reason, but you
1:05:50
have to make sure all of it works in your network well because you might be
1:05:53
swapping one vendor out for another. I mean, there's all kinds of challenges
1:05:56
inherent to that, too. Um, we don't have 50 vendors, so that makes it simple.
1:06:00
Our, like I said, the the way our network is is designed, it can
1:06:04
accommodate all of our vendors, so we can plug and play as needed for growth.
1:06:08
Um, and that's an advantage we have, but it is not easy. So, I don't have an
1:06:12
answer. The let's talk about like 12 months.
1:06:18
I don't I don't know if we'll see terabit in 12 months, but we're going to
1:06:21
see terabit pretty quickly. as you know capacity is just going to the moon at
1:06:27
this point in terms of demand. Mhm.
1:06:30
>> What is the timeline for rollout going to be going forward for increase in
1:06:35
capacity and backbone capacity? Right? Because um I don't think a lot of people
1:06:39
understand the actual you know pricing dynamics that goes into you know like a
1:06:43
network like yours upgrading right. So if you just look at at a pop at a one of
1:06:48
these carrier hotels you know there is um there is an
1:06:53
inflection point where it is cheaper to purchase and do lags. You know, we're
1:06:57
saying 4x400, right? Like there is cost and power, cost and line card, cost and
1:07:02
optics, right? Is it cheaper to do 4x400 or 2x 800, you know, like and and with
1:07:11
every new, you know, every time this comes out, like it's cheaper to do 4x400
1:07:15
for a while and then it becomes finally cheaper to do 2 by 800. Yes. or you need
1:07:19
more capacity, you kind of get forced into 4 by 800, you know, and so there's
1:07:24
it's not this like, oh, 800's available, let's just go roll it out. It's like,
1:07:27
okay, you know, 800's a, you know, available. When does it hit the price
1:07:30
point that it actually makes sense for us to switch, right?
1:07:33
>> It feels like that window is starting to shorten for a lot of networks in terms
1:07:36
of the capacity demands that are being pushed onto these networks of like old
1:07:40
math isn't working, isn't mathing anymore, and and that push for
1:07:43
performance has has overtaken it.
1:07:46
>> Yeah, it's a great question. I don't have an answer to it. You know what I
1:07:48
can tell you is is that the the migration from 10 gig to 100 gig was
1:07:53
fairly gradual over the course of several years. You had COVID in there.
1:07:57
You had some other things that delayed some of that. I mean it it spiked up
1:08:00
capacity but the to your point the move from 100 to 400 has been dramatic is
1:08:07
very short shorter than the previous jump. And and so you know to solve the
1:08:13
problems you got a couple different options you can do. one by you know one
1:08:18
400 or do four 100s and it all depends on where and how fast and all that
1:08:22
stuff. So we have all those conversations with customers this so we
1:08:24
can accommodate the capacity but how we how we solve it could be a couple
1:08:28
different ways depending on where it is. Um, we haven't rolled out 800 yet. It's
1:08:33
tested in the lab and it's working. But to say what it what the price and the
1:08:38
the capacity of the impact from a timing perspective is going to be, I don't have
1:08:41
that answer today.
1:08:42
>> But I have a I have a fear that it's going to be even faster than uh 100 to
1:08:48
400. Uh I find a lot of companies when they
1:08:53
start looking at doing any sort of bonded links that haven't h I mean again
1:08:57
it's like this is a weird thing where it's like you know depending what your
1:09:00
lineage and your experience and background with it is there's just
1:09:02
certain things you don't like don't you don't know until you run into it right.
1:09:05
>> Yeah.
1:09:06
>> If you're doing 4x100 links you know your capacity for a single stream is you
1:09:11
know whatever your interface is you know you've got a you've got a 100 gig limit
1:09:15
in terms of how fast you can move traffic. Yes, you've got 400 gig worth
1:09:18
of capacity, but you know, one one stream can be 100 gig. And trying to
1:09:23
work through that in capacity planning is also interesting with people. It's
1:09:25
like, well, do you need just capacity?
1:09:28
>> I'm glad you said we're not going to get into the technology too much in this
1:09:31
>> a little bit. We're going deep in there now. We're talking capacity plan,
1:09:34
>> but we're talking capacity plan, but we're talking but it but it applies,
1:09:37
right? Because now you start talking about it's like you know I mean how much
1:09:39
is your sales team actually going back to your customer and saying hey we want
1:09:42
to buy 4x4 you know are the is it does the order cycle start with we want 400
1:09:45
gig of capacity and you're like do you want 4x100 do you want 1x400 you know
1:09:50
they say we need 800 are you saying do you want one by 800 2x400 you know 8
1:09:54
by100 like how that because that again has a different cost implication to you
1:09:58
in a data center you have crossconnects you have to account for right like
1:10:01
that's usually not your side of the table but
1:10:03
>> No it is it definitely is but I think I think it really you know um there's no
1:10:08
one answer fits all and in some places we can accommodate the requirement
1:10:12
easily. Sometimes we have to say look we have to move these cards from here to
1:10:16
here to accommodate this or we can only do so much at this this this uh pop you
1:10:21
know so there's a lot of there's a lot of conversations that take place about
1:10:24
how we configure the solution that they need. It typically starts out with a a
1:10:28
capacity requirement and then we say this is how we can fill it and here's
1:10:31
the path you know that we would do it in and they say nope we want to change it
1:10:35
do this and this and this. It's rarely that they say we want one option, one
1:10:38
way and that's all we'll take.
1:10:40
>> It's always a conversation. So again, that's really a conversation between the
1:10:44
customer's technical team and my technical team on what we can fulfill in
1:10:48
what way and what they can live with. And there's a lot of considerations in
1:10:52
that decision. From a sales process, you you're going to get to a point, you
1:10:55
know, if somebody's ordering a 10 gig, you probably don't have a capacity
1:10:58
conversation, right? At 100 gig at this point, you're probably not having a
1:11:01
capacity conversation. But at some point, you're getting to a you're
1:11:03
getting to a point, right, where You have to have a capacity and you know
1:11:07
>> yeah n by 400 we got right we need six yeah
1:11:10
>> so what you know how what is that process like in the sales in the sales
1:11:14
cycle right somebody's talking to you and say I want this you know whatever
1:11:18
right and how long does it take to go through a capacity plan how long is it
1:11:24
you know do you see before you're saying okay great no problem we have this we
1:11:26
can do it we can engineer it we can we can it's redundant we know we can you
1:11:31
know have network health with this you know and then at what point you get into
1:11:35
a conversation it's like okay we can build this capacity for you but it's
1:11:39
going to take this much time because we have to touch these many boxes in
1:11:42
between point A and point B. It's all of those things and nobody's in a position
1:11:45
where they can say we got I mean 10 gig yes but when you're talking multiple 400
1:11:50
gigs across multiple you know areas of the world geographies of the world um
1:11:57
we're doing just that you know we we put a business case together and we say
1:12:01
we're we're good here we have to do we have to expand this here we have to
1:12:06
build here and we go back to the customer with that because you can't you
1:12:10
know you can't fake you you either have it or you don't and you can get creative
1:12:15
and you can do alternate paths in some cases but um alternate routes but that's
1:12:20
not what they want because you have to be adverse from this or they have a
1:12:22
certain latency that is going to be caused because you took a different path
1:12:26
than what they had. So you know it's it's definitely a conversation and you
1:12:31
just have to be as transparent as you can with customers because nobody can do
1:12:34
everything. You know the other dynamic that I've seen play out a lot. I mean
1:12:38
this is again two decades worth right the slower circuits have less cons I
1:12:45
don't say consumption less less actual traffic flowing across it right so your
1:12:48
average 10 gig circuit at this point you know even if it's IP transit and they're
1:12:50
doing like a one or two gig commit on a 10 gig circuit that circuit probably
1:12:54
isn't using more than 3 gig on average right
1:12:56
>> especially in the enterprise phase.
1:12:58
>> Exactly. Right. But then when you get into the fast circuits like those fast
1:13:01
circuits aren't running at 10 20 30% capacity. somebody that shows up and
1:13:04
says, "Hey, I want to buy 400 gig from you." It's like, "Oh, you're going to
1:13:07
use 400 gig from me."
1:13:08
>> Nine times out of 10, yeah. No, and that that all comes back to the planners and
1:13:12
and how you run your network, how hot you run your network, because we have to
1:13:16
be, you know, there's there's uh committed rates that they say, we'll
1:13:21
commit to this, but we want to be able to burst to this, if you will. So, you
1:13:24
have to accommodate for not only committed, you have to accommodate for
1:13:26
the burst across the big. So, it's not a simple conversation, but I would say
1:13:32
when you're focused on doing one thing
1:13:34
>> Mhm.
1:13:34
>> for 30 years, that's where that benefit I say comes in. It's not easy, but we
1:13:40
don't have to worry about what's going on in the voice world or the SDWAN
1:13:42
device, you know, or anything else. We we really focus on that. Um, and our
1:13:47
planners are amazing, you know, and we we we've got folks that have been with
1:13:50
us for 15, 20, 30 years in some cases, and they really understand the network
1:13:55
and how it works and what we can and can't do. So, I put a lot of faith and
1:13:59
reliance on those on the network teams because they, you know, anything we we
1:14:04
bring to the table, we say this is a big build or this is a big opportunity. How
1:14:08
do we address this? Well, a lot of times it's we can do it this way. Go back to
1:14:11
the customer and say yes, we can do it this way. And they go, great. let us say
1:14:14
we can do this this way, but this other this other aspect we we have to build or
1:14:18
or do something. Go I'll go a different direction. I'm okay with that. You know,
1:14:22
I I don't I don't know if there's any secret sauce or anything to it. Um I
1:14:27
don't know enough about the network planning all the cycles they go through,
1:14:31
but I put a ton of trust and faith in those guys. And again, it's all we do.
1:14:35
So,
1:14:36
>> what do you wish people knew about Aurelion that they don't? Right now, I
1:14:40
can I can I can give you a couple different leads into this, right?
1:14:44
Because there's, you know, like the old old network guys like me have been in
1:14:48
data centers for a long time that have known you for a really long time, right?
1:14:52
So there's like
1:14:54
>> what am I not thinking about, right? Like my audience. But then there's also
1:14:58
the enterprise, right? Where the enterprise has not interacted with you
1:15:00
probably ever or even know about you. Like so outside of like just knowing you
1:15:04
exist, right? like what would you like you know in in very disparit audiences
1:15:10
you know what should people know about you well at the end of the day I would
1:15:13
say that we differentiate ourselves or solve three specific problems in the
1:15:18
market right the first I've touched on quite a bit is network performance um
1:15:22
we've never strayed from our knitting what we do well for the past 30 years
1:15:25
we've not tried to be all things to all companies we've been very focused on
1:15:29
network you know on connectivity if you will and so I think for me having a a
1:15:35
ubiquitous homegrown network spanning the globe is a definite differentiator,
1:15:42
but it's not just talk. It's you you see it in the network performance. The other
1:15:47
thing that that um that I I I see that we don't really get credit for or talk
1:15:52
about is that we not only provide a great service, but the delivery
1:15:56
experience is much better. And a lot of the conversations we have with
1:16:00
enterprises and and some wholesale companies, wholesale companies tend to
1:16:04
leverage multiple providers, enterprise one or two. And so they're terrified
1:16:09
because the delivery experience they had getting to the point where they are was
1:16:12
horrible. And there's a risk at making a change.
1:16:16
>> But one of the things that makes us great is our delivery experience. And I
1:16:20
wish people more people knew about it. And you know, a lot of my competitors
1:16:23
will will do um they'll do a centralized delivery organization. and they'll put
1:16:28
it in uh uh Manila or in in a lowerc cost country and they'll work 24/7. But
1:16:34
the challenge you have there is that you know
1:16:38
>> the guys that deliver they deliver and they see the green light and they say
1:16:42
any problems after that hand it off to customer service and you you're working
1:16:46
with a a a support organization that is probably very good but they were not
1:16:50
part of the solution discussion. They don't know what's been delivered and
1:16:53
they're starting off, well, give me your circuit ID or, you know, back to and and
1:16:57
that's not where an enterprise needs to be when they're having a delivery
1:17:01
challenge,
1:17:01
>> right?
1:17:02
>> At Aurelion, first off, we have a a globally distributed uh delivery team.
1:17:07
So, I have teams in all different time zones. They speak 26 different languages
1:17:11
within that team. We understand the ecosystem, the customs, the time zone
1:17:16
challenges, all those things. And usually the companies that we leverage
1:17:20
for tail circuits and things are also customers, right?
1:17:22
>> So they've worked with the provider that's providing the tail, you know what
1:17:26
I mean? So we have such a better delivery experience than most because
1:17:31
we're not all centralized. We're globally distributed. We're working in
1:17:34
the time zone at the location where that IT team is. And then the second thing is
1:17:39
is that we hold on our service delivery team is part of the solution design
1:17:43
conversation. there's a handoff of knowledge that they hold on to that
1:17:47
delivery until it's working as advertised in the customer's
1:17:50
environment. So, they're not handing you off to some known name in another
1:17:54
country who's not been a part of the journey in the other seven locations. We
1:17:58
hold on to it.
1:17:59
>> It's it's a huge differentiator because it's all fun and great when we sign the
1:18:03
contract and everybody's excited,
1:18:05
>> but the real work starts in the delivery and that's the first taste of a customer
1:18:08
to your company.
1:18:09
>> Yeah. And that delivery experience deters companies from making changes.
1:18:15
>> So, so what you're saying is you're a little different from the phone company
1:18:18
and if they're experiences in working with the phone company, it's a little
1:18:20
different.
1:18:21
>> It truly matters and we recognize it's a risk and we bring our delivery groups in
1:18:25
on sales calls because people that's a real concern of enterprises to say this
1:18:30
is how we're different. And then finally, I think the big different
1:18:33
another the third problem we solve is customer support. I know everybody says
1:18:38
that, right? Um, but you can measure it in very tangible ways. So for us, you
1:18:43
know, our net promoter score is 70 and that's one of the the the highest in the
1:18:48
industry. And we get a lot of good great comments about, you know, the sales team
1:18:51
is great, the engineering is this, the delivery was this, all the things I just
1:18:54
said are are we're showing in our in our net promoter score. But the other thing
1:18:58
we have is that we we have a little less than 1% churn, you know, of our customer
1:19:04
base. So, we've grown t- dramatically over the past few years in revenue.
1:19:10
We've grown faster than anybody in the market, but we've grown and I think part
1:19:13
of that is is because we don't have to worry about what's going out the back
1:19:16
door. Yeah.
1:19:17
>> Right. When you're not churning now, customers downsize and they change. So,
1:19:21
BK, we count all of that as churn. Yeah.
1:19:24
>> But we're still at a very low churn rate. So, our customers typically stay
1:19:28
with us for, you know, at least three or four term uh term cycles. They don't
1:19:33
leave us. they typically grow and and uh they say great things about us, our net
1:19:37
promoter score. So, by any measurement stick, we're one of the best support
1:19:41
companies in the business. So, it's it's the network performance. It's the
1:19:45
ability to deliver at a very high level and and and then support those customers
1:19:50
long term to be a partner in the business, not just say, "Hey, we see the
1:19:53
light. Good luck to you. It's on your side." You know what I mean? I think
1:19:57
those are things that I wish more knew about us because we get a lot of open
1:20:00
doors. we're doing very well, but I mean there's, you know, we we don't sell to
1:20:04
everybody. We sell to a subset of a subset. If you don't have a global
1:20:07
backbone, we're not that interesting to you, right? But if you're there, I think
1:20:11
we're the best fit out there. Um, I always say we nobody's perfect, but we
1:20:16
definitely do more right than we do wrong. And I'll put my company delivery,
1:20:20
performance, and and uh support against anybody in the industry. as somebody
1:20:24
that has a as a team of people just help with implementation of client projects.
1:20:30
I can tell you what's what I companies that go my favorite one right
1:20:36
you know it's like working with a data center when you have an infrastructure
1:20:38
in a data center and then they somebody will order a circuit from a le and be
1:20:42
like oh we've ordered the circuit for the le and le's like okay we've
1:20:44
installed in the empo and you're like that does nothing for me because I need
1:20:47
it in the data center
1:20:48
>> yes
1:20:49
>> and then you're like oh well this particular building has multiple data
1:20:52
centers in it which data center controls the empo that the le is in
1:20:56
>> and then how do you get you know and like things that that on the surface you
1:21:01
just have these moments where you're like this is such a insane problem to be
1:21:04
working with right now and how much time and aggravation gets lost in those
1:21:09
things. And so, you know, just as like a practical example to what you say in
1:21:13
terms of like having experience and working in these markets and
1:21:15
understanding these things, like there is a big difference between a company
1:21:18
that has experience working with data centers and carrier hotels and companies
1:21:24
that don't and also how do you do capacity planning for people that I mean
1:21:28
like it's just
1:21:29
>> it's just completely different, right? You know, you're it's not apples to
1:21:32
apples here, you know?
1:21:33
>> No. And and I think the enterprises as they get more into AI and they look at
1:21:37
more of the wholesale like services to take on these big traffic logs like an
1:21:41
IP transit like an EVPL I I think that's really where the rubber meets the road
1:21:45
right I always joke we never lose twice we might lose once because you know
1:21:50
somebody bought the business they went low ball and they made a financial
1:21:54
decision but the experience between the implementation the ongoing support was
1:21:58
terrible
1:21:59
>> yeah do they were they able to install I mean you talk about capacity planning I
1:22:02
was in a Um I was in a peer group um a long time ago and
1:22:08
uh you know the the the host was asking people you know what do you what what do
1:22:12
you want different from us right and they were talking about specifically
1:22:15
like interconnections and and in carry neutral facilities and and one of the
1:22:19
network engineers that was there was like I want an API to your system to to
1:22:23
request circuits and get an auto LOA and send that to the blah blah blah and get
1:22:26
a cross connect just turn this whole thing up instantaneously. and you're
1:22:28
like, um,
1:22:30
>> you can't there's there's reasons why we can't do that for you, you know, like
1:22:34
>> I I I know you'd love just to be able to go out and be like, hey, I want to, you
1:22:36
know, at that time it was um it was it was 10 gig, you know, 100 gig hadn't
1:22:40
rolled out yet.
1:22:41
>> And you know, and and like like of course you're like that's that'd
1:22:45
be great, you know, like again the promise of the cloud like you can just
1:22:48
instantly turn on, but like everybody wants to deal with a cloud company, but
1:22:51
in our business people to physically move things and touch things still.
1:22:54
>> Yeah. you're like, "No, somebody has to go there and plug stuff in, you know,
1:22:57
like or somebody's digging a trench and putting stuff in the trench."
1:23:00
>> But to our earlier conversation, that's the attraction of a network as a service
1:23:03
sort of value proposition. But anyway,
1:23:06
>> I um look, I'm I'm a I'm a technology
1:23:11
maximalist. I, you know, I believe that that technology improves the human
1:23:15
condition. Agreed. And the and the more that we get, the better off it is. And
1:23:19
you know, it's worth the, you know, boy, I'm going to get hate for this one, but
1:23:22
like nobody's complaining that we don't have people operating elevators anymore
1:23:26
for us, right? Like new buildings don't have buttons in elevators anymore,
1:23:30
right? Like, you know, so
1:23:33
>> the experience is better as a whole. And and so I'm always I'm always I always
1:23:39
love to see it. What what I wish that we saw more of in technology wasn't just
1:23:43
wasn't this rush to pretend like we're marketing against
1:23:47
you know like what's the new hotness right
1:23:50
>> but more like how are we actually solving a problem what do you need oh
1:23:53
well we've decided that we need GPU capacity right where's our data or our
1:23:58
data is here GPUs are going to be over here how do we now talk make them talk
1:24:02
to each other why is this important to us oh because we're doing this to you
1:24:06
know
1:24:07
>> for our R&D for for you know, for for data science, for revenue, for you know,
1:24:13
like all those, you know, like what's the actual thing that you're trying to
1:24:16
do? And you're like, okay, connectivity isn't that sexy on the surface, but then
1:24:19
you're like, well, I mean, guess what? You know, like this is the glue that
1:24:22
holds businesses together.
1:24:22
>> Well, you bring up an interesting point. You mentioned API, and I think that, you
1:24:26
know, yes, that's kind of the nature of our world these days, right, is how can
1:24:30
I have that cloud experience, right?
1:24:32
>> But you also have a a a a new breed of decision makers who that's how they want
1:24:37
to deal with it. They want to deal with an app. They want to deal a text. they
1:24:39
want to, you know, they don't want to talk to somebody or have a meeting with
1:24:42
someone. And so, I mean, one of the things that we're pushing is a lot of
1:24:46
self-service in our in our portal,
1:24:48
>> but but it goes beyond just, oh, you can put a treble ticket in or you can do
1:24:53
this BGP testing and you can do this uh physical, you know, uh, test or
1:24:58
whatever. It it's definitely a toolkit for diagnostic, but it's also we're
1:25:02
rolling out other tools. And so what I mean by that specifically is um we have
1:25:07
an internal tool, but we're going to be exposing it to our customers that you
1:25:10
can go in and it has all of our network map on it. You can go in and look at our
1:25:15
network map and say, you know, I not only I want to buy away from this point
1:25:18
to this point and I want it to be this capacity. I want this. I want that. You
1:25:24
can actually do that and you can download that and send it into us and
1:25:30
say, "Price this out."
1:25:30
>> Right.
1:25:31
>> Right. Well, we're evolving that to where if you have a price book with us,
1:25:36
you can get the price and you can actually go in and reserve that
1:25:39
capacity. Now, you'll have to sign the contract within a certain it doesn't
1:25:42
hold for at nauseium, you know, forever. But we are looking to say how can we
1:25:46
automate not only the service portion of the portal, but also uh the network
1:25:52
exploring what can and can be provisioned. Uh and then we also when
1:25:57
you reserve that capacity we give you a delivery time and some if it's an
1:26:00
existing can be you know matter of a week or two.
1:26:03
>> That's awesome.
1:26:04
>> And if it's if it's not we say nope we have to go back and do this or do that.
1:26:08
But we're looking to do as much self-service not just on the support
1:26:11
side but also on the sales side so we can streamline those conversations to
1:26:15
get down to I built I built this network. I just need you to price it and
1:26:21
you know how fast can you deliver? I
1:26:22
>> I I feel like you've been holding out on me here. This is this is a actually a
1:26:26
really interesting thing to talk about, right? Because now
1:26:32
there's a side of that which is great and there's a side of that that's bad,
1:26:35
right? Because now you start talking about like
1:26:39
people don't know what they don't know, you know, and if there's not like any
1:26:43
sort of like thought process be some be, you know, in in some like loops like oh,
1:26:48
you know, people can get them dig themselves into holes without realizing
1:26:51
that they've dug themselves into a hole because
1:26:53
>> the the idea of it isn't to to automate the human touch and the knowledge out of
1:26:59
it. The idea is to lower the friction of doing business with companies and say if
1:27:04
we can give you some tools where you can play around a bit and build something
1:27:08
you think is interesting. The next step is let's validate all this with a
1:27:11
solution engineer to your point because you may be making some fundamental
1:27:14
mistakes based on your the rest of your network or your traffic pattern or your
1:27:18
application environment or your cloud. You know what I mean? But it's it's to
1:27:21
lower we want companies to say it was so easy to work with these guys. And part
1:27:25
of that conversation is giving them some a toolkit and giving them the ability to
1:27:30
go and design some stuff their own where they're not locked into talking to
1:27:33
somebody if they don't want to. There's a lot of smart IT people in in the world
1:27:37
and they just they just need the ability to
1:27:39
>> I had no idea you guys were offering this. I got we got to yell people here.
1:27:42
>> If you it's not it's not commercially available today, but it's absolutely the
1:27:45
direction. But if you go to our website today and you go into our interactive
1:27:49
map, you can go highlight a region and you'll get all the pops and all the pop
1:27:54
addresses on our in our network and you can download those and use however you
1:27:59
want to. We have a map you can download. You download it in Excel or CSV file or
1:28:04
whatever you want to. That's the first iteration of it.
1:28:06
>> So we're going to be expanding on that sort of, you know, you can go back to a
1:28:10
region. I want to do Europe. you can get all the pops in a certain region and
1:28:14
then we're gonna say you can go and do build a build a wave or build you know
1:28:19
certain services there.
1:28:20
>> So let's go let's go really basic here. We'll keep it with Ethernet. We won't
1:28:22
talk about like waves and optical just say Ethernet. And you know you have two
1:28:26
core products in Ethernet. You have some sort of private connection.
1:28:29
>> Call it whatever you want to call it. And then you've got a internet
1:28:31
connection right? Call it whatever you want to call it. Different people have
1:28:34
different terms trade terms.
1:28:35
>> Um like traditional you'd order an internet
1:28:40
circuit. You say, "I want this much capacity. I want this kind of contract
1:28:43
on it." You get a port into your network, right? Like boom, we connect
1:28:46
and I have an internet circuit.
1:28:48
>> Or I say, "I want a private connection." Same thing. I go through a provisioning
1:28:51
process and we get a port.
1:28:53
>> Assuming that it's connecting to the same gear on your side, right? Like one
1:28:56
does one thing, one does the other thing. Part of, you know, I network as a
1:29:01
service. Part of what, you know, you see these like NAS products are really doing
1:29:05
is just being able to deliver multiple service types on a single physical port.
1:29:08
>> We can do that. If you look at our website, we're already there. You can
1:29:12
have one one port, multi-ervice, single port is what we call it. So we already
1:29:16
do that today. So within that then you start talking
1:29:21
about provisioning and provisioning capacity across your network. If I've
1:29:24
got a port in here and a port there and you're doing your cloud onramp, you've
1:29:27
got this all integrated. So I can say I want this cap capacity to this region of
1:29:30
this cloud provider and I want it delivered on this, you know, Q tag in
1:29:34
this market and boom, just turn on.
1:29:36
>> Yes,
1:29:37
>> you guys are burying the lead on this stuff. Yeah, you know, I mean, you
1:29:40
always have to things move so fast. You always have to validate that capacity is
1:29:44
available. You always have to validate the card inventories there. All those
1:29:47
things. So, yes, but these are things that we've been doing for a while. So, I
1:29:52
think I think it's uh
1:29:54
>> it's it's well known in the wholesale space, not so much in the enterprise
1:29:57
space.
1:29:57
>> Yeah. I I would say probably not well not known at all, you know, because this
1:30:01
is I think this is the core promise of a lot of what you see is like upstart NAS
1:30:05
companies, you know, and and and really it's like that flexibility is like
1:30:10
product service delivery and then contractual flexibility, right? You can
1:30:14
go to a NAS company and you can say I want to buy, you know, consumption based
1:30:16
utilization across your your network.
1:30:18
>> Well, today a lot of the functionality is internal,
1:30:21
>> but where that translates is we have a quick install time, faster provisioning,
1:30:25
all those kind of things. But the goal is to expose more of that to the enduser
1:30:30
community in a controlled way. When you start putting pricing out there, it gets
1:30:33
a bit dicey, you know, what you do.
1:30:35
>> So, how do you how do you ring fence that to where existing customers can see
1:30:39
their price book or you know and versus somebody coming in off from the internet
1:30:43
or whatever.
1:30:44
>> So, um we're not completely launched in all that today, but the first iteration
1:30:49
of that is out on interactive map and we'll continue to launch over the next
1:30:53
year. There's a road map. So, so the goal is to, like I said, reduce the
1:30:57
friction. And if people want to buy from us directly, make a phone call, we'll do
1:31:01
that. If they want to buy through the channel, we're great with that. Go
1:31:04
through the channel. If they want to buy and they go and they want to do it all
1:31:07
online, we're great with that, too.
1:31:10
>> It's a I I mean, I I I appreciate the complexity of this, right? I don't think
1:31:16
people really understand. I mean, you know, you've got a thousand plus devices
1:31:20
on your network that you have to manage and orchestrate, right? And then when
1:31:23
you start saying like okay not only you have a thousand plus devices on your
1:31:26
network you've got you know an order of magnitude more network interfaces on
1:31:30
your network and then you've got you know uh BGP you've got internal routing
1:31:35
protocols you've got cloud connectivity requirements that come from the
1:31:37
different providers that's a very complex you know configuration set and
1:31:41
now you're like oh you know on top of that we want to have some sort of
1:31:43
automated API based delivery you know process that goes into orchestration
1:31:46
that has like appropriate guardrails on it that doesn't like break something
1:31:49
like whoops you know um So I I I'm I'm like old school in the sense that like
1:31:56
my priority is like don't break anything for me, you know, like first and
1:32:00
foremost like like
1:32:01
>> I I will tolerate like having to interact with humans in exchange for not
1:32:05
having stuff break, right? But um you know but it's it's it's it's
1:32:11
like you know you know thinking back again like going back 20 years to like
1:32:14
now and like what's available not just from like the performance but like this
1:32:18
conversation you know like my brain is like oh I can
1:32:22
solve that problem here and here and here now. Yeah.
1:32:23
>> Yeah. Yeah. Well I mean again I'll I'll take it back to where we don't do
1:32:27
everything.
1:32:27
>> Yeah. We do a few things really well and we've done it for a very you know big
1:32:32
audience in the wholesale space but now as AI is permeating more than just you
1:32:37
know wholesale and hyperscalers are engaging with more than just you know
1:32:42
carriers and big companies all over the place. Um we definitely want to want to
1:32:48
stand out a bit and shine and say look we're we're here and these are things
1:32:51
this is why we're here today doing these podcasts to get the word out there.
1:32:54
Okay. So, so, so parting parting words, right? We're we're at an enterprise.
1:32:57
>> Oh, are we recording? I didn't know we were.
1:32:59
>> Yeah. Yeah. Yeah. So, partying partying words here. Um, yeah. I do it sneakily
1:33:02
so that way we can just kind of walk in. Um, you know, 30 second elevator pitch.
1:33:07
We we are in an elevator at a conference together and it's like, oh, you know,
1:33:10
hey, what do you do? Right? And and not a wholesaler, not an air carrier
1:33:14
customer, not I'm not a hyperscaler, right? Like, you know, I represent like
1:33:18
a a large mid-market enterprise, right? Fortune,000.
1:33:23
What's what's your what's your response to that?
1:33:25
>> I I kind of already did it, but I'll do it again. So So you know, Aurelian's one
1:33:29
of the largest connectivity providers in the world globally. We really solve
1:33:32
three main problems, whether it's wholesale and enterprise. We we we
1:33:35
specialize in in global network performance. Uh we've always stayed true
1:33:39
to a lot of my competitors have uh tried to be experts at SDWAN and security and
1:33:44
voice and what they've done is really dilute their value proposition where
1:33:48
they're experts at nothing. We really have stayed close to just providing
1:33:51
connectivity services on a global scale. As a result, we have the best network,
1:33:55
the highest performing network and the amazing reach that we manage day in day
1:33:59
out. It's our main course of business and we we feel that's been the served us
1:34:03
well especially in the days of now we're seeing AI volumes drive and cloud
1:34:07
volumes and all the different things that are happening in the ecosystem.
1:34:10
We're in a very good spot. But secondly, it's really about that delivery
1:34:14
experience. My competitors have a uh centralized delivery. you know,
1:34:19
organization, typically not in the country where they're based for cost
1:34:23
reasons. Um, because that you get somebody that sees a green light come on
1:34:26
or they they think it's installed and they say any problems, go to customer
1:34:29
service. That's not a great experience. Conversely, Aurelian's got a globally
1:34:33
distributed uh a globally distributed uh delivery force. They they're all over
1:34:38
the world. They understand the languages. They understand the cultures.
1:34:41
The players involved. Those companies that are providing those tail circuits
1:34:44
or anything else are typically our customers. So we have a very different
1:34:47
relationship with the people providing services as part of the solution than a
1:34:50
lot of my competitors. And they're on the ground. They work in the time zones
1:34:53
that your offices and data centers are in. There's a tremendous value of that.
1:34:57
But most importantly, they hold on to that delivery experience until it's up
1:35:01
and running as designed in the customer's environment. We might pull in
1:35:05
some engineers. We might pull in some support people to help us, but you're
1:35:08
never going to be tossed off to another organization who doesn't understand the
1:35:11
solution or it's been delivered thus far. And then finally, it's about the
1:35:15
support. You know, delivery and support are the two biggest things why customers
1:35:19
leave. Cost is always in there, but but delivery experience, they don't forget
1:35:23
about that. And support, it's always it's constant. Um, by any measurement
1:35:27
stick, Aurelian provides the one of the best uh support experiences in the in
1:35:31
the in the business. Be it net promoter score, be it our low churn numbers, the
1:35:36
fact of what our customers say about us, they continue to grow and our just
1:35:40
growth as a company. We're growing faster than anybody in the in the
1:35:42
business. I think, you know, we we do it right. I always say we do more right
1:35:46
than we do wrong. We're not perfect. Nobody is. But in terms of global
1:35:50
network performance, having an amazing delivery experience, and being being
1:35:55
really responsive and supportive along that journey, I think that puts us in a
1:35:59
different position than most of my competitors today.
1:36:01
>> Amazing. Thank you very much for the time.
1:36:04
>> Hey, sure.
1:36:04
>> I had so much fun. This was this was fantastic.
1:36:06
>> Thank you. I appreciate it.
1:36:09
>> That's it for this episode of Signed. If you got something out of this, share it
1:36:13
with someone in your world who's staring down a tech decision. A CIO, a CFO, a
1:36:16
founder, a procurement lead, whoever. That's how the show grows. Everything
1:36:20
from today lives at itbroker.com/mpodcast.
1:36:23
Show notes, transcript, links to anything we mentioned. If you're in the
1:36:26
middle of a real tech decision right now and you want someone in your corner
1:36:29
without the vendor bias, that's what we do at itroker.com. Schedule a call on
1:36:33
our website. Buy tech without regret. I'm Max Clark. Thanks for listening. See
1:36:38
you on the next one.

‍