Today’s episode features an interview between Matt Trifiro and Joe Zhu, Founder and CEO of Zenlayer. In this interview, Joe discusses Zenlayer’s evolution from edge data center provider to the goal of becoming the world’s #1 edge cloud provider, as well as Zenlayer’s focus on being a truly global company and its unique capability and passion for serving emerging markets.
Today’s episode features an interview between Matt Trifiro and Joe Zhu, Founder and CEO of Zenlayer.
Joe founded Zenlayer in 2014 after 8 years of managing global business for ChinaCache, China’s largest CDN provider.
In this interview, Joe discusses Zenlayer’s evolution from edge data center provider to the goal of becoming the world’s #1 edge cloud provider, as well as Zenlayer’s focus on being a truly global company and its unique capability and passion for serving emerging markets.
Key Quotes
“It’s always been my ambition to become a global company, because my passion is to improve the global user experience, not just individual regions.”
“Our choice is to go to emerging markets, so-called underserved markets like India, Indonesia, Vietnam, Russia, Brazil, where the infrastructure is not as good as in what we call mature markets like the U.S., Europe, Japan, Korea. But the user base is huge and their mobile users have dramatically increased year over year, and they have a bigger population as well as a GDP increase.”
“Even though our strategy is to go to the emerging markets first, our customers need global coverage. Because the experience is not just for one location. It's going to be for global users.”
“Obviously we don't know what's the next killer app yet. There will be a new killer app coming online. Is it a different type of e-commerce or different type of social application or different kind of a live streaming? Just imagine in the future, for example, you and I sitting in the same room, like a projection. But that requires a lot of bandwidth.”
Sponsors
Over the Edge is brought to you by the generous sponsorship of Catchpoint, NetFoundry, Ori Industries, Packet, Seagate, Vapor IO, and Zenlayer.
The featured sponsor of this episode of Over the Edge is Zenlayer. Improving user experience doesn't have to be complicated or expensive. Zenlayer helps you lower latency with on-demand edge services in over 150 PoPs around the world. Find out how you can improve your users' experience today at zenlayer.com/edge
Links
[00:00:00] Matt: Hi, I'm Matt Trifiro I am the chief marketing officer of vapor IO and also the cochair of state of the edge. And I'm here this morning with Joe Zhu, the CEO of Zenlayer. And hi, Joe, how are you doing today?
[00:00:11] Joe: Hey, Matt, how about you?
[00:00:13] Matt: Oh, I'm, I'm doing terrific.
[00:00:14] So I wanted, I wanted to start out by asking you, you know, how, how you got involved in technology. Like, what is it that drew you to technology?
[00:00:23] Joe: Yeah, it's a long story. So actually, well, I studied at MIS back in the college, you know, and I literally love internet everything, but at that time, the dialogue. Okay.
[00:00:36] So I wrote a couple of software for the companies. The one is actually I wrote a, a scheduling system for China, Eastern airlines. That was they, they actually, that's their first scheduling system.
[00:00:49] So yeah, but fast forward, when it come to us, I started bill in the run, their global network of phone for one of the largest carrier. That's I, where I get to [00:01:00] the internet and the technology.
[00:01:02] I literally
[00:01:03] Matt: about, what, what year was this approximately?
[00:01:05] Joe: it's a 2000. It's a bright after the internet first bubble.
[00:01:09] Matt: crash. Right. Okay.
[00:01:13] Joe: Yeah. Build a global network.
[00:01:15] Matt: is actually a good time to get into the internet. It turns out.
[00:01:17] Joe: It is. It is, you know, every scene is so exciting. Right. So, yeah. So I build it like a data centers, global pop, everything. Yeah. That's how I started.
[00:01:29] Matt: Yeah. That's great. And, and how long have you been at the head of a Zen layer?
[00:01:35] Joe: Literally, I founded a company in 2014. Yeah. Right, right. After I left the China cash. That's it? China cash is the one over the CDN company. Yeah.
[00:01:45] Matt: Okay.
[00:01:46] Joe: was on 14 now it's over five years.
[00:01:48] Matt: And so, so were you, when was the transition from engineering to business management?
[00:01:54] Joe: That's a good question. You know
[00:01:56] that
[00:01:57] Matt: Although I suspect you're never, you never let the [00:02:00] engineering go completely. It sounds like.
[00:02:01] Joe: Oh yeah. I'm still engineering by trade. I mean, every day it's still, I still need to touch the computer and you still need to log into the Laura Robbins, you know? Yeah. So let's let engineers, you can fix the network and you can build a network, but later part of my.
[00:02:18] Well, when I worked for a previous company, I had noticed I can probably make an even bigger impact on the business, say, where's my technology background. That's, that's why I started kind of like have the transition from engineer become a, a more of a, technology dealership in a sense, right? Yeah.
[00:02:38] Matt: Yeah. Yeah. That's great. And so you founded Zen layer in 2014, is that right?
[00:02:44] And so back then, and the world's changed quite a bit, and I want to talk about how the world's changed, but, what, what problem were you trying to solve?
[00:02:51] Joe: I try to, it tried to solve the issue at the edge, right back in 2014, public cloud is already [00:03:00] huge. Right. But how to input the user experience? It is a, is a big question. So this, this is a challenge and an opportunity for us. So I start work on this. Yeah.
[00:03:13] Matt: Yeah. And so when you say improve the experience at the edge, like what, what, what in particular, like what, what customer problems, what, what, where did you see the opportunity to, who do you want? Who do you see the option to solve for first?
[00:03:24] Joe: A meeting entertainment. So the interactive experience, right? Like e-sports, there's like a 2014. He, sports is getting hot. Right now is obviously
[00:03:36] everywhere. Right. But how to make a million users to compete as on the internet. That's a big issue.
[00:03:44] Matt: Yeah. So that's really interesting in 2014, that was your pretty early, you know, so, so when you look at like the history of edge computing, the first reference I can find to edge computing actually came out of the team that founded Akamai when they were at MIT. And as you know, that version of edge computing was one [00:04:00] way.
[00:04:00] It was essentially a, a data cache. Right. and it made a drastic improvement in how the internet functions, I mean, videos used to buffer and websites took forever. but 2014, you know, there weren't a lot of people that were thinking about edge computing on, from an interactive standpoint on the return trip.
[00:04:17] So, I mean, what, what made you, I mean, I guess the growth of e-sports, but I think it's a pretty big. Insight, I think in 2014, to recognize that you want those edge computers doing more than just, you know, cashing and doing small modifications to webpages. What, what, yeah. What, like what, where did that insight come from?
[00:04:37] Joe: Yeah. So before I funded, generally I run China cash global business for seven years.
[00:04:44] That's a CDN company, right? So, so it's the static caching. So obviously CDN solve the videos and the, the static content delivery issue. But obviously that's several years gave me a lot of insights. There's [00:05:00] lots of accustoming.
[00:05:01] The interactive experience those in experience, it's still those issues still not addressed,
[00:05:08] Matt: Yeah. Even today in some senses. I mean the, yeah. Yeah. So, so I mean, who are your first customers?
[00:05:16]Joe: the first customer actually for Zimmerli it's I don't know, it's on the record now, but I can tell you
[00:05:22] Yeah.
[00:05:24] Matt: Really.
[00:05:25] Joe: Yeah. So, so the first cosmic it's, obviously we have a longterm relationship backing days was a Google, but, it was a Google needed deploy the service and service in China.
[00:05:39] first, China is a challenging market. Very interesting market for this way.
[00:05:45] So, yeah. Yeah. So we were helping them to deploy a city's edge locations and fully managed and managing the equipment service as well as technology means, routing everything, make [00:06:00] sure that the eight cities able to cover a whole child.
[00:06:05] Matt: Yeah, well, and I noticed now that you're off, you have an offer a barrier. Bare metal cloud offering. And it, it, it's interesting that that was kind of your first experience. They're building a cloud, helping Google extend their cloud,
[00:06:17] into these edge data centers.
[00:06:19] Joe: you kind of learned that through the process, how the big guys do doing right then you, you want to kind of build a, a subset that can open up for every company can use it.
[00:06:29] Matt: Yeah. And so now, it seems like you've evolved your business. You're offering a full line of services. Why don't you tell me a little more about like Zen layer today.
[00:06:39] Joe: Yeah, I'm anybody today now is we evolve defy edge data center provider to an edge of cloud poet means we fully automated compute network in the, in the application service compute, we are offering the bare metal cloud. Eventually we will move up the stack. For example, like the offer [00:07:00] contingent service, a function as a service.
[00:07:02] Okay. So those are a, we are on these work on it. We will offer that very soon. Right now. What can you do is since we, we have more than 150, the global edge data centers at that time, the first initiative was a connect all status for ourself. Then we figured our customer actually walked that connectivity
[00:07:26] Matt: So you built a backbone that connects all your data centers together. And of course your customers that are in both of your data centers are gonna want to use your connectivity.
[00:07:34] Joe: Exactly and they'll be virtualize it so you can pay as
[00:07:37] you go in and adjust a
[00:07:39] Matt: Did you do build that, that SDN software yourself?
[00:07:42] Joe: Yes. So initial phase, the first first version is a base are open for long. We are working on the second. We call SDN 2.0, it's actually based on tele MACI and the second Marathi. So. It's complete different, different technology.
[00:07:58] Yeah.
[00:07:59] Matt: Interesting. [00:08:00] and so, so you said a hundred, 150 data centers worldwide, 150 edge locations. And so, so let's, let's think about, about, so how, how has your concept of edge and the needs that, that have emerged on the, in our, in our are still emerging for applications. How's that changed? What you.
[00:08:25] Joe: yeah, 2014 is that at that time, it was still very heavy on static contents. Even we, we see the trend is the interactive, the, experiences coming. Right. But now, today is really heavy leap, interrupting experience. I think, CBN aesthetic content was still there still play a very big part of the digital experience, but is more of a move to the other trend because user's behavior has changed.
[00:08:54] Right.
[00:08:55] Matt: Like, what Like, what are the, what are the big changes in user behavior?
[00:08:57] Joe: Yup. So it's very [00:09:00] fragmented by the user's interests. The sending or receiving information is very fragmented, not like a million users interested in one thing. Now it's like a group of a
[00:09:10] hundred users,
[00:09:12] Matt: should in a million things.
[00:09:13] Yeah, that's good. That's a
[00:09:15] good way of looking at it. That's interesting.
[00:09:17] Joe: And it's a very global, you have users from Jakarta and in Los Angeles or Mexico, they want to compete.
[00:09:25] Or do some social scenes like today, if there is a call social watching. Right. But you and I can watch the same movie.
[00:09:33] Matt: Yeah. Right, right. Yeah. It's these plugins to your browser. That'll synchronize your Netflix and yeah, that's pretty cool.
[00:09:39] Joe: Right. So all those are intact interactive experience versus in the past is a static experience.
[00:09:44] Matt: Yeah.
[00:09:45] I mean, I've got two kids and they're entire social world. exists online and it's live and interactive, you know, with their, with their characters, but also voice and video and, yeah, it's a very, very different world. Especially when you look at the young generation. [00:10:00] Now there's a, there's a, there's another trend.
[00:10:03] And I wonder if, if Zen layer. Or, has, an interest here. I'm sure you probably do, which is, you know, you talk about consumer behavior. You're talking about the, the humans talking to machines and one of the big trends that I see that's
[00:10:15] driving a lot of this
[00:10:16] business is machines talking to machines and, you know, humans, humans have.
[00:10:21] Tolerances of latency and the ones of seconds machines once a, once a second is glacial. Right. So I'm wondering how, how that is changing, how Zen layer, is thinking about its its business and its relationship to, to
[00:10:36] providing edge compute services.
[00:10:39] Joe: Yeah, for sure. I think that that's the future, right? And not only human to talk to human, human talk to device, device, to device, right. That's the Indy to any connectivity. So today we only doing a little bit research on that. The reason being that is we don't know what's the business model for that business.
[00:11:00] [00:11:00] When the device talk to the device, IOT is already big, but the traffic level is very low. Great. So for us, obviously,
[00:11:10] exactly right. How they intacted each other, we don't know yet. Alright. So what we were going to do is that the next phase of mini in two or three years, we will offer SoCal IOT managing suite on our edge network.
[00:11:27] So we help her to manage it. Our team, not necessarily taking improve the user experience here. When does the device start talking to each other? And we will try to adjust that experience,
[00:11:40] Matt: Yeah. And that, and that's interesting. I noticed, on your website, I think, maybe on your blog, you have an ambition of globally reaching. A large percentage of the population with 10 millisecond latency and 10 million seconds is amazing, right? Like that's really fast for a human.
[00:11:55]and so when you project into the future and you imagine this IOT service that you [00:12:00] want to launch at some point, what, what kind of latencies do you think you'll need to, you know, To compete with sort of on prem solutions or to approximate to deliver, deliver something at the edge.
[00:12:10] I mean, it's 10 millisecond sufficient, or do you, do you think you're going to have to even get lower latencies?
[00:12:14] Joe: Yeah. So first I think always service all intentions, never going to compete with imprint because this is a different use case right now, back to the 10 milliseconds. I think a soror research is like a, for autonomous driving or remote matters medical use cases make, I think a 10 millisecond should be sufficient.
[00:12:36] Hey, because that's the farthest away. We can the goal because you want us to get 10 millisecond, most likely in the future, we need to deploy edgy cloud service
[00:12:48] inside the tower.
[00:12:49] So close to the
[00:12:50] Matt: well, I mean, that's, my company is building those data centers. So when you have that need, we can, we can certainly accommodate you.
[00:12:56] Yeah. I mean, where our goal is is, 75 [00:13:00] microseconds to the Radiohead.
[00:13:02] Joe: Yeah, exactly. Then from the radio head to the user, that's the last we, we,
[00:13:06] we can't, there's nothing we can do,
[00:13:08] right?
[00:13:09] Matt: Yeah. Although the, the, the, you know, certainly 5g, will improve that last mile. And, you know, today, fixed, Y you know, wired, fixed networks, didn't achieve that. And, and I'm actually seeing some demand to shift on prem into these, across a low latency network. You know, it's sort of this, this emerging near prem.
[00:13:26] Opportunity. So it's interesting how these worlds are blurring. In fact, you know, one of the things I, I was pretty inspired by your, your open letter to 2020 open letter you published in early January. And one of the things that you wrote there that I thought was a really cool way of thinking about it, and I'm just going to read it, read it out loud to the audience, which is, you said, quote, as the boundary between digital and physical worlds, increasingly blurs the demand for real time, interactive digital experience has exploded.
[00:13:52] And I think that, that, you know, you, you actually imagine. The physical and digital worlds, you know, merging. And you think [00:14:00] about like that's, what's happening to cloud is coming closer to the edge and the, the edge or the devices are coming closer to the cloud. And there is almost this actual mixing of the physical digital, but also this kind of metaphorical mixing the digital, I thought that was a really eloquent way of describing, you know, kind of broadly why the edge is so.
[00:14:19] It's important to us now, why we're talking about it and you know, maybe why it wasn't that important, or common topic,
[00:14:25] you know, 10 years ago in 2010,
[00:14:27] right?
[00:14:28] Joe: Exactly. Exactly. And especially once five G's really come online. The last mile speed is dramatically increased, but the latency and the bandwidth is there. So the user do this demand for, for better experience will be. The pressure will be huge. So, but as you know, the telecom carrier is never able to build a mint mile fast enough.
[00:14:54] Matt: Yeah, that's right. That's right. Yeah. Yeah. That's very true. so one of the things that I, that [00:15:00] I, I picked up in, in, in researching gen layer and talking to some folks in your company is that you have a particular focus on, emerging markets. And, you know, I, I, I think I'm hearing two things. I think I'm hearing, that, that is definitely where you've got traction today, but I'm also think I'm hearing, you've got much more broad global ambitions.
[00:15:18] And I'm wondering if you could talk about, The importance of emerging markets, how you serve them uniquely and then whether or not I'm right in picking up that, you know, you see your market expanding beyond that.
[00:15:29] Joe: Got it. Yeah, because it's a, it's a choice, right? Because for a startup you have limited resource of funding talent. So our choice is a week go to the emerging market. So-called underserved the market
[00:15:43] Matt: So, so let's describe for the audience, doesn't know like, like T talk to me about what you define as an emerging market and how the internet is different there, and then how you help the internet
[00:15:55] Joe: Got it. Got it. Yeah. So in our view is, like, [00:16:00] India, Indonesia, Vietnam, right? Russia, Brazil. So the infrastructure is not as good as we call mature market, like a U S and Europe, Japan, Korea, but the user base is huge. And then we also look at those markets. They're mobile user subscriber has been dramatically increased year over year, and that has a bigger population as well as the GDP increase.
[00:16:28] Matt: Yeah. I mean, that's the primary source of internet access is mobile, right?
[00:16:31] Joe: Exactly. They don't have broadband in the past. Right. So they just jumped through that generation, go to mobile internet. Right. So for us, it is since it's a bigger market industry, served in our technologies really can help them to improve the experience. Why not? Yeah. Then if we can do really well and approval a business case, and as well as technology, then we can go back to the mature market [00:17:00] because a mature market, it's a well-served and there's a really highly competitive,
[00:17:04] to be honest
[00:17:05] with them.
[00:17:05] Matt: Yeah, right. Well, I mean, it's right. As a startup, it makes sense to go after a niche where, where you feel like you can invest, more creatively and differentiate yourself. so, who are you serving? I mean, who are you? Who are your customers? You know, you say, look, we've got this, this improved experience that we can help you deliver in these, in these other countries that are underserved.
[00:17:26] And maybe you weren't able to deliver the kind of experience that you wanted to, but now we can help you deliver your service. I said, who are your customers? Who's looking to, to do this. I mean,
[00:17:34] besides Google.
[00:17:37] Joe: So our customers are mainly obviously to be a B to B business and it's media entertainment. So just naming a film, for example, tick tock. Nope. So even it's emerging market doesn't mean it's the users seeing as USD. They use tech tech every day. Right. The [00:18:00] thing is for gaming company, like a mobile ledger.
[00:18:02] Well, one of the very large e-sports company.
[00:18:06] Matt: Right? So,
[00:18:10] So, your business is B to B, but most of your customers are B to C.
[00:18:14] Joe: Yeah. That's right. Yeah.
[00:18:16] Matt: it.
[00:18:19] Social media or yeah, video sharing, that sort of thing. Yeah. That's, that's really interesting. so your value proposition is, Hey, you want to expand and improve your service in these markets? We have a
[00:18:31] press, the easy button.
[00:18:32] Joe: Exactly right. It's a, it's a, I mean, put it this way. You pay for the expense.
[00:18:40] Sorry to see that. But
[00:18:42] Matt: Yeah. Yeah, yeah, yeah, yeah. That's right. Well, and I'm sure it's a lot cheaper than them having to go in and figure out, you know, the, the, the deployment scenario and all these different markets. Cause you're right. And you mentioned, you know, your first customer taking into these edge locations in China, like some countries it's
[00:18:57] hard to do business.
[00:18:58] And just the fact that you [00:19:00] figured out how to do business in reserve rooms, a lot of the barriers, plus you've, you've. Figured out where the data centers are and whether you've had to build or buy or lease or whatever. And so you've kind of, you've made it and, and they can always in their head think, well, we could always go in ourselves, even though that may, you may create a really sticky service.
[00:19:16] Yeah. That's really interesting. Now I noticed on your network map that you have data centers in emerging markets, you know, you've got one in Los Angeles and San Jose. How does that fit into your, your business today?
[00:19:26] Joe: Yeah. So, I mean, even our strategy is to go to the emerging markets first. Right. But how our customer need a global coverage. So lucky us where I've taught 12 data centers and that we would love to partner with. You guys could feel more like a 50 or hundreds, right. Because the experience is not just for one location, mate.
[00:19:48] It's going to be a global users.
[00:19:50] Matt: yeah,
[00:19:51] yeah. And so do you, do you find that, that when you approach a customer, if you say, look, we have. let's just say complete coverage. That's, you know, the definition [00:20:00] of Alyssa, but yeah, we've got complete coverage. You go to one company, we can get you to the entirety of Brazil or the entirety of India.
[00:20:05] You find that that's sort of the,
[00:20:09] Joe: Yeah. So either a region or country specific or they say, okay, you're just helping me improve the global user experience that it doesn't matter which country is.
[00:20:19] Right. So we'll just help you.
[00:20:21] Matt: Yeah. Yeah. That's, that's, that's, that's really interesting. and T tell me about, know, where you saw the opportunity around, I mean, it's, it's one thing to offer the capability to help, you know, Google bring its own servers into a country or anybody tick talkers. And it's another thing to say, we're going to buy servers and we're going to deploy them, and then we're going to lease them to you.
[00:20:44] And we're going to start looking like a Google, at least a small one, right at the start. How did you go through that transition to the bare metal cloud offering and, and, you know, as you say, moving up the stack, that whole strategy to move up the stack.
[00:20:56] Joe: Yeah. So, I mean, when we started was helping other people [00:21:00] to deploy, right. But the value for us, a startup, where we keep a single well, how to continue to create value. If you just pay pervade, providing the data centers in the managed service, the value is somewhat. Is it a little bit low, especially when the customers are getting more sophisticated, right?
[00:21:22] So, and then we can only help certain customers, but now we stopped building the stack, so called edge cloud infrastructure service. Then we will continue move to the, profit as a service means not only large companies in a small companies and developers, they can do. They also can deploy a service, deploy the code at the edge.
[00:21:45] I think it will be, make an even bigger impact. And it's a good business case. Right? So
[00:21:52] that's, that's our choice is to continue to move up. Right.
[00:21:56] Matt: Yeah. are you seeing a demand for more bespoke hardware and, [00:22:00] I guess what I mean by that is, is, you know, so in order, if I want to run more sophisticated AI workloads at the edge, I might want to have bare metal with GPS or, or there's some specialized FPG AEs that are being used.
[00:22:12] And are you seeing demand for like innovation on the equipment that, that,
[00:22:18] that they can get through your bare metal cloud? Tell me a little bit about that. What you're saying.
[00:22:21] Joe: Yeah, so for sure, because there is a demand, put it like a GPU or more at events, the hard way in, at the edge that the issue is rhino is just cost and the benefit is not there. Right. Because if we have a clients running, cloud-based the gaming. So, you know, edgy sites, each server, they want a pupping for AGBU.
[00:22:49] That's not a issue. The issues of where's the power. You don't have enough power. I'm finished it in Melbourne. But [00:23:00] as you say, you have 20 racks. If you popping bump bunch of GPU power, that service.
[00:23:05] You probably can only put two racks in there. Right. So
[00:23:09] Matt: Well, there's creative ways around that, but yeah, yeah, no, you're right. I mean, I mean, you know, whether you can actually rectify and cool the equipment, is one question, you know, whether you have the facilities to do it. then there's another question which is like, can you literally get more power to the site?
[00:23:24] You know? I mean, it's, it's, you know, a couple of hundred kilowatts is doable.
[00:23:29] Joe: Yeah, that's right. But obviously I think eventually this will happen, but it takes a while. Is it hard? Like how is like a Nvidia maybe come up a low power consumption GPU? I
[00:23:43] don't know.
[00:23:44] I'm just
[00:23:44] Matt: sure. I mean, you look at the work that arm's doing and the,
[00:23:49] Joe: Exactly
[00:23:49] right. When is what we always think about a software at a software layer? Basically, basically I have a 10 edge sites in the U S for example, [00:24:00] point GPU is all in those 10, as you say, it's somewhat, I make it a synchronized.
[00:24:07] Matt: Yeah. That's interesting. Yeah. no, that's valuable. Right? Cause otherwise I'd have to build up myself. So you offer that as a service, the sort of
[00:24:16] Joe: Yes, Brian though, we were not on the GPU side. This is a really far away, but on the bare metal side,
[00:24:22] on the compute side, we kind of make that experience available now.
[00:24:26] Matt: Yeah, so let's list. Okay. So 10 data centers in the U S and you described them as ed data centers. So what makes those data centers edge data centers, as opposed to just
[00:24:35] an in market regional data center? What, what is it that differentiates that?
[00:24:39] Joe: Yeah, 10 data centers is not really correct myself. If I deploy 50 or a hundred a year, your dad has said that that's edge, that these 10
[00:24:55] Matt: Yeah. And I, and I think the regional data center is part of the edge, but it's that, it's that sort of [00:25:00] outer layer and there's this, you know, there's, there's gradations there. And as you said, you know, it's like you said, like right now, you're not seeing a lot of demand from your consumer based customers for.
[00:25:10] Superior to 10 millisecond. And so there's a lot of workloads that can operate in that 10 to 30 millisecond range. and frankly, as you say, like it's, it's, it's less, it's more cost effective to deploy in larger facilities that have more power that, you know, it's. And so I think there's a, there's always going to be a role for these regional data centers.
[00:25:28]so essentially you simply see talking about the number of data centers, and I think that's probably a reflection of how far out in the edge you are. I mean, it's reflection of the global coverage, but also how far on the edge you are. So if you're at 150 data centers today, you sort of, you know, fast forward to 20, 25, and let's assume that most of your ambitions are realized how many data centers do you think.
[00:25:45] Thank you. You'll be in there.
[00:25:51] Really? Yeah. I would say it might be over a thousand,
[00:25:55]Joe: that.
[00:25:59] Yeah. [00:26:00] Yeah. You know that the difficulty is when you have thousand data centers, it's just imagine that it was, you have million servers in one data
[00:26:09] center, like an Amazon
[00:26:10] or a million servers in solid bed.
[00:26:12] I said,
[00:26:12] Matt: Yeah. Yeah. Yeah, it is. It is. And I think that, that, you know, you talk about the software tooling that you've built to do these simultaneous deployments on your belt or metal servers like that. That, I mean, you know, when there's, when there's two places, you can provision an ECE, two instance, U S Western use like a human can make those decisions.
[00:26:32] When there's 15 human can make those decisions, but when there's 50 to a thousand, we're going to need a whole new set of orchestration tools. And the is probably going to change by the second, you know, cause I think, cause the resources is going to be constrained and you're probably going to be buying instances on auction and it's going to be a really interesting, interesting world.
[00:26:54] And we're just gonna get a lot of software tooling to make that work and you can see how it's, where it's going. I [00:27:00] mean, to some extent. Companies have solved this inside the data center. I mean, that's what, that's what Google's Borg and coop, the Kubernetes derivative of that does. and you can, it's a logical extension to say, well, you know, I, if I can treat the 40,000 cores in a single data center as kind of one it's pull of resources, why can't I spread that across a hundred data centers,
[00:27:22] but we're not there yet. We haven't yet got all
[00:27:24] the orchestration tools.
[00:27:26]
[00:27:26]
[00:27:26] yeah, that's what I'm. what do you see as the emerging customers? Like, like when you see, like the, you know, the customers that your, that your bread and butter customers today, that you've got a clear value proposition. so you mentioned it's the, it's the, it's the e-sports let's, let's talk through those.
[00:27:40] Let's talk about Zen layers customers today. Like
[00:27:42] who, who, who are your key customer segments today?
[00:27:45] Joe: So gaming gaming.
[00:27:50] Matt: So when you said gaming, you mean both like e-sports, but also like, just, you know,
[00:27:57] Joe: Yeah. As long as the interactive [00:28:00] a gaming experience.
[00:28:01] Yeah. Those gaming companies, right. And the online education, how long has your kitchen is a huge,
[00:28:07] Matt: with the global pandemic conditions.
[00:28:10] Joe: exactly. And I'll as you occasion is, you know, we have a customer, is that. The teacher is all us in Dallas or in, in Montana. But the students that are in Asia, like in Shanghai or Beijing or Jakarta, right.
[00:28:25] They're learning English. So that millisecond of delays really, really important.
[00:28:31] Matt: Yeah. Well, so, so tell, tell me how that works. So I'm a teacher in Dallas and I have, I have a, a student in Shanghai. And let's say I'm not as only, or customer what's that experience like, and then when I turned Zen later on, how does that change and how has it changed
[00:28:47] from the experience, but also like technically what happens?
[00:28:49] That's different.
[00:28:50] Joe: Got it. So if you're not using any service providers like us, you're pretty much a Dallas teacher connect to the [00:29:00] Dallas is through the internet, connects with Dallas pup. In a sense, then it goes to the public internet, which Shanghai, the student art was also connect to Shanghai's internet. Like yeah, China telecom se obviously you can imagine that experience.
[00:29:18] There's so many components that will make it stop working. No use our technology today is that that Dallas teacher will connect to our bare metal cloud
[00:29:30] Dallas.
[00:29:31] From there, our private backbone will take over
[00:29:34] Matt: I see. So I'm not BGP routing over the traditional internet. I'm I'm going over your backbone directly.
[00:29:42] Joe: you, you pay for a fat, fast track.
[00:29:45] Matt: I see. I see. So, so yeah. Oh, that's interesting. So you're you're yeah. So I can see why that's a, a big part of your business offering this to these sort of global networking services. And so it's not, it's not magically, you know, solving for the speed of light.
[00:29:59] It's just [00:30:00] reducing the network hops and, and giving you sort of, you know, higher qualities of service across your routers. Yeah. Yeah. And that, that means there's less delay in the video, less delay in the audio list.
[00:30:11] Yeah. That's, that's really great. Okay. So, so education, what else?
[00:30:15] Joe: Yeah. So hold on to the education yet. So in the future for the education companies, they don't need to care about bare metal cloud and they're all cloud networking backbone anymore. They can just call an API. So means their application.
[00:30:34] Just send us the requirements that I want to 120 millisecond between Shanghai and
[00:30:40] Matt: So just like, I, I create a declarative
[00:30:42] manifest for my Kubernetes workloads. I create a network
[00:30:48] Make it happen or tell me, tell me I'm out of luck.
[00:30:53] Joe: it and the scale up and that's it.
[00:30:57] Matt: That's very cool. I know. And I can see, I can see why that, that would [00:31:00] be
[00:31:00] a real pain to build myself yeah.
[00:31:04] Joe: Yeah. Okay. So online education e-sports we talk about larger ceremonies or live streaming. All right. Tick tock. Oh, all those. Now there's like a high school seniors. They're using tech talk and use a chat or a posters show videos. Right? That, that, that obviously requires a lots of edgy service,
[00:31:28] traditional CDN service as was that you confused?
[00:31:33] yeah.
[00:31:34] Matt: Yeah. That's really interesting. Are you seeing any, you know, B2B applications like with, with business SAS software or anything like that?
[00:31:42]Joe: yes. So right now, work from home. So when they access to like office three 65, even zoom, right. Especially globally, when you have a meeting globally or you access the SAS application globally, there's a challenge. That's why they will provide us so-called [00:32:00] application
[00:32:01] service, SAS as a
[00:32:03] service, which helping them to improve the application experience.
[00:32:07] Matt: Yeah.
[00:32:08] Do, do you, do you think that your, your background having come from China and now being in the U S
[00:32:16] Joe: Yeah, it's always my ambition to become a global company because it's not a Chinese company and just the global company, because it, my ambition, what am I? My passion is to improve the global user
[00:32:31] experience, not individual regions.
[00:32:35] Matt: Yeah. Yeah. That's that makes a lot of sense to me. you know, one of the things we touched on earlier was, you know, the difference between what, what is needed now,
[00:32:46] and what's going to be needed
[00:32:47] in the future.
[00:32:48] And I think another way of saying that is like, there's some, you know, irrational, exuberance around edge, and there's like the practical, like what actually needs to be done today.
[00:32:58] And can you give our listeners [00:33:00] some, some insight into how, how you're seeing these trends evolve and where you're seeing, where you think the new demand is going to emerge? you know, even if it's not here today, you
[00:33:10] know, what's, what's coming for, like what, what are the, what are the macro trends that are.
[00:33:15] From your perspective that you're seeing in your, your, your customer base
[00:33:19] and others that you're interested in.
[00:33:20] Joe: Yeah. So obviously we don't know which what's the next killer app yet. We are mimicking today. What we do, like for example, we have video conference, then we think about next, maybe next two years or three years that we have AR VR based. Video conference, right? So w w we will mimic your, while we doing today, but still that process, well, Queenie a huge demand.
[00:33:45] Right? Then later it will be a new killer app we're coming online. We don't know, is this a different type of e-commerce or different types of social application or different kind of a [00:34:00] live streaming, right. But. Just imagine in the future, you and I just like sitting in the same room. Right. But the projection is there that requires a lot of bandwidth.
[00:34:13] Right.
[00:34:14] Matt: Are you,
[00:34:15]
[00:34:15]when do you think, when do you think that'll be here?
[00:34:15] Yeah, you're you're, you're, you're, you're waiting for it. Yeah. I'm surprised that AR VR hasn't yet. Taken off. cause it's just, to me, it's an amazing experience and I just, I, I, I don't understand. So I think it's inevitable. I'm like you, I just think it's inevitable. Maybe, maybe we just have to get away from the heavy goggles and you know, other sort of apparatus.
[00:34:31] Joe: Yeah, I think a bandwidth is your shoot. The hardware experienced is the issue and content is the key. There is service.
[00:34:45] Matt: ya, I'm been locked up my house for three months and I just would die to walk around Paris virtually. Yeah. Or even Los Angeles. Right. Well, yes. Right. Wouldn't that be? That'd be [00:35:00] cool. You know, that's one of the things that I've discovered, you know, and I've, I've been working at home for, for many, many years, but now that everybody's working for him, I think, I think people are starting to realize this, that I think there was this, this assumption that human experiences are so dramatically different in person that you can never replicate it with with technology,
[00:35:20] but. W what I've discovered from my own experience is that great communication like this conversation, like it's like watching a movie, we just get lost in the movie. Like I have to think about the fact that we aren't sitting across the table sharing a cup of coffee. Right. And, and I think the part of what makes that possible, it is the richness of the experience.
[00:35:43] This is video it's real time. It's got voice. I mean, the listeners don't know that we're actually looking at each other over, over a zoom call, but that's how this podcast is recorded. and I just notice it disappearing and I, I. I can only imagine when, when we retuned the experience, [00:36:00] you know, with other modalities, like three dimensional space, how it's it is going to be a new, you know, like my children.
[00:36:07] My children project their lives online, like to them, you know, you know, I remember, I remember when I showed my youngest child, the Mac for the first time he tried to touch the screen. Cause that's, that's, that's what the phone does. Right. That's what the iPad does. And I think there's just this expectation, this next generation that just like.
[00:36:25] Our human interactions are enriched and expanded by all this technology. And it's not a replacement for human interaction, but it's a, , it's, I'm finding it a very profound experience. And it's neat to see this, you know, kind of
[00:36:38] technology nobody ever sees. You know, I mean,
[00:36:40] how many people actually see Zen layer,
[00:36:43] Joe: Yup. Yeah,
[00:36:44] I think
[00:36:51] Matt: Yeah. Yeah, it's a, it's a really, so, so, you know, when you think about, you talk about like, you know, improving the internet globally, right.
[00:36:59] As being one of your [00:37:00] passions, what, what makes you passionate about that? Like you really light up when you, when you say that, where is that drive coming from?
[00:37:06] Joe: I just want to make people happy.
[00:37:08] Matt: That's awesome.
[00:37:10] Joe: Yeah.
[00:37:10] Matt: Really? You just want any people having
[00:37:14] that's a, that is a really awesome
[00:37:16] mission. Wow. I got to go try that one on that's a, that's a really good one. That's a really good one. this has just been a wonderful conversation. I mean, is there, is there anything that we should've covered and didn't, or that you'd like to, to, to tell the audience or, you know, because this has just been a delightful conversation.
[00:37:33] Joe: Thank you. I think that, yeah,
[00:37:41] Matt: Awesome. Well, thank you, John. I appreciate you being on the podcast and, you know, it's it's been wonderful meeting you and I hope we get to meet in person someday.
[00:37:49] Joe: Yeah. Yeah.
[00:37:51] Matt: Hey John. We're,
[00:37:52]
[00:37:52] we're wrapping. [00:38:00] Oh, I said, John, did I say John? Oh my gosh. Okay. So what I tell you? Oh my God. I just totally. Yeah,
[00:38:07] Joe: Yeah.
[00:38:08] Matt: So I need to say, did I only do that one? So. Okay. So you want me to say, what do you want me to say? Just thank you. Thank you for Joe or it's great. It was great talking to you. yeah, yeah. Thanks again, jail. That was really great talking to you and I hope we get to meet in person someday.
[00:38:30] Joe: Thank you, Matt. Great. Looking forward to.
[00:38:33]