Today’s episode features an interview between Matt Trifiro and Jacob Smith. Jacob is a co-founder at Packet and currently serves as Vice President of bare metal strategy and marketing. In this interview, Jacob discusses the founding of Packet and the vision for bringing the world of software into the world of physical networks, cloud infrastructure as a craft, what he sees as the next stages of edge computing, and much more.
Today’s episode features an interview between Matt Trifiro and Jacob Smith. Jacob is a co-founder at Packet and currently serves as Vice President of bare metal strategy and marketing.
Jacob co-founded Packet in 2014 with the goal of democratizing hyperscale infrastructure capabilities. Since then Packet has been one of the early leaders in edge computing for business, and earlier this year was acquired by Equinix.
In this interview, Jacob discusses the founding of Packet and the vision for bringing the world of software into the world of physical networks, cloud infrastructure as a craft, what he sees as the next stages of edge computing, and much more.
Key Quotes
“We're inviting more people in, and especially with the edge, it's a huge opportunity to bring the world of software into the world of physical networks that currently power the internet.”
“The idea of the buyer being different, software getting way big, infrastructure getting more specialized-- this all led us to think, ‘How do we get out of the way? How do we focus on delivery model? How do we focus on fundamental things?’ And that's really what we set out to do. That's what led us to the edge.”
“I'll bet on software anytime. I think that properly served up access, innovation will occur. There are a lot of innovation-minded people touching internet infrastructure.”
“Instead of guessing the use cases, I think it's better to look at how to be more open. Look at what the clouds have done super well-- they've created ecosystems. I think we can create the same mindset-- this diverse, totally Wild West, weird, special world called the edge.”
Links
Connect with Jacob on LinkedIn
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 Packet, an Equinix company. Packet is the leader in bare metal automation. They are on a mission to protect, connect, and power the digital world with developer-friendly physical infrastructure and a neutral, interconnected ecosystem that spans over 55 global markets. Learn more at packet.com.
[00:00:00] Matt: Hi, I'm Matt Trifiro. I'm the chief marketing officer of Vapor IO and I'm also the co chair of the state of the edge project of Linux foundation. And I have with me here today, Jacob Smith, the chief marketing officer at Packet, and he is also my co-chair at the state of the edge.
[00:00:14] So we've known each other for quite awhile. we'll try to not use too many inside jokes and, Talk, all things ed. So Jim, how you doing today?
[00:00:22] Jacob: I'm doing great. it's a beautiful day out in the middle of nowhere. I'm calling in from Southern Vermont, which is known as the crappiest part of the internet in case you're wondering. So if my, if my voice breaks up, it's not me. It's just the copper wires connecting us.
[00:00:37] Matt: Yeah, you're up there deploying some edge data centers. Is that what you're doing to
[00:00:40] improve the, the video conferencing?
[00:00:41] Jacob: Yeah. It's the obvious like, capitalist endeavors to go save, you know, and bring a little latency gaming to the 12 people who live near me for sure.
[00:00:53] Matt: Well, it is great that you have such a, a job that gives you so much mobility, and maybe we will come back to that. But before we [00:01:00] start talking about edge and current day, I'm super curious to know how you even got into technology. I mean, as I understand it, like you have an opera. Background a master's degree in music.
[00:01:14]probably play three or four instruments. Maybe you could cut a well, let's talk about that a little bit. So what instruments do you play?
[00:01:20] Jacob: Okay, cool. The only one that I play well, I wouldn't even say now, but I do how to do it. It's the one thing I know how to do well is the bassoon. So when I was young, I started, playing the saxophone cause you know, Debbie Gibson and all these great, like eighties pop songs that had saxophone in
[00:01:37] it. And, and so exactly right. And so I asked for a saxophone for my birthday when I was like 10. And, even though that was great, cause the saxophone is like literally meant to be easy to play. And so I was feeling great. but my brother and sister were playing in the orchestra. And what do you now? They don't have any for [00:02:00] saxophones and the orchestra, except for like once a year.
[00:02:03] And so I was like, well, how do I get the new orchestra? And they were like, well, no, one's playing the bassoon. And I was like, sign me up. I didn't ask, like, why is no one playing the bassoon, ends up it's super complicated and, you know, keys to your left, thumb and forefinger, right fond. And those all have to work together.
[00:02:20] And blah-blah-blah so I started playing the bassoon and really fell in love with it, you know, just because it was complicated and mechanical and. You know, it was like Mike early hacking. And so I, I started playing the bassoon in high school and then just kept going. So I went to undergrad at Carnegie Mellon, which is like super technical school, but I went for music and I spent most of my time hanging out in the computer lab because it was open 24 hours.
[00:02:48] I mean, can you imagine
[00:02:49] Matt: Where were you? The resident, the soon player in the, in the computer lab,
[00:02:54] Jacob: well, Carnegie
[00:02:55] Matt: hat, a hat, a guitar, a bassoon case for tips.
[00:02:58] Jacob: Well, it was really funny [00:03:00] because, Carnegie Mellon was a very interdisciplinary school. And so you could do all these kinds of cross collaborative, classes. So I took one called building virtual worlds. Which required someone from the music department, the art department, computer science, like they made everyone, you had to fill up enough so that you could actually create a world.
[00:03:18] And so, you know, that got me into working on Silicon graphics computers in 1998 and it was super fun. I was like, wow, this is neat. I was already into computers. Cause I grew up kind of in that. What do they call it? The Oregon trail generation, you know, right. Between analog and digital, you know, go to radio shack when they still had it.
[00:03:38] And, you know, like making my own computer because that's what you do when you like, when people don't make computers for you. Right. So
[00:03:47] Matt: Well, I think those of us with liberal arts degrees get some, some ribbing from our, our technical, twin brothers, for example.
[00:03:54] Jacob: for instance,
[00:03:54] instance,
[00:03:54]Matt: but I tend to see, I mean, so. Music has a lot of mathematics in it [00:04:00] and mathematics and technology. Obviously it go well together so I can see the connection.
[00:04:06] Do you remember the first computer you ever owned?
[00:04:08] Jacob: yeah, absolutely. So my brother and I, my twin brother was so we were in business together. So you might hear about that more. but he and I had a killer logline business. We charged $10. a week and then $30 a month on subscription, like, cause you gotta get people on recurring revenue. Right?
[00:04:27] Matt: When your kids, I
[00:04:28] Jacob: Exactly. And it was the best business because my parents covered all the capital equipment, you know, and we just make cash. So we saved up $2,000. It was a lot, but that's what got you. Like. Two gigs of Ram. And, you know, I can't remember the clock speed, but it wasn't very high. You know, this was Pentium.
[00:04:52] Matt: So, is it a PC, a dos computer or windows? Computer? Yeah.
[00:04:57] Jacob: Absolutely. My aunt had worked for [00:05:00] IBM. And so she had helped us buy when we were younger, a P S junior, which we had in the house. And that was super fun growing up with. But of course it was not. The scene is like when you could start going down to Fry's electronics and start figuring out how to build your own, tweak it out.
[00:05:16] And so that's, that was the first computer that we built. and it was, I don't know, we were 12.
[00:05:22] Matt: Yeah. Wow. You've come quite away. So, so you're the chief marketing officer of packet. also one of the co founders. tell me a little bit about, about when you found it a packet and why you found it, what you saw the opportunity being.
[00:05:35] Jacob: Yeah. Well, it was mainly driven by a personal opportunity, which was, my brother had been, he actually went to Juilliard as a musician and then got into tech right away. So in 2001, he got into tech, and started working for an early Linux hosting company
[00:05:52] he played the double bass. Yeah. So he, yeah, so we were both on the lower end.
[00:05:59] Matt: You both checked the wrong [00:06:00] box on me. What
[00:06:00] instrument you want to play for him?
[00:06:02] Jacob: Exactly. And so he, he had sold the company in 2011, into the internet, which is a, you know, a collocation and cloud company. And he was looking for something to do. so he spent a year or two kind of floating around and, and I was also kind of ready. I was playing in the opera, like you mentioned earlier.
[00:06:19] I was playing the soon I was, I was working in nonprofit and I had a marketing business on the side. So it was doing everything. And I was hostile. I had my side hustles before they were popular and we were like, well, you know, maybe we should do something together. And the rural was anything, but infrastructure infrastructure is a big boy's game.
[00:06:37] Now this is like, this is not the way it was in 2001, 2002. You know, you have to have a lot of capital. This email, the cloud thing is like legit. so around 2013, we started looking for something to do, and we actually tried on a couple of different business ideas. one was, we were going to do a blockchain based or backed by a.
[00:06:56] Real gold. That was, that was really cool. [00:07:00] and we started a couple other ideas, but none of them really made sense because we didn't have any access survey, anything special to add to it. And so, finally we had this, kind of a ritual of hosting that. A holiday party in New York, every December and inviting like all the old voxel crew and all that, you've just, you know, just the friends.
[00:07:19] Right. And everyone was all working at different places, going and working at tech companies and other things. And we sitting around a table in the basement of the museum for Chinese and America. And it was a whole bunch of network nerds going like, you know, this isn't the way it should work. We should have EGP for everyone.
[00:07:39] And this layer two stuff it's for the dogs. And they were just like, I didn't understand the words they were saying, but they all seem very passionate that whatever the way was going now was the wrong way. And the better way would be to do it a different way. And I was like, We should do that. And that's really what started us down the path of infrastructure.
[00:08:00] [00:07:59] And then we worked on it for a couple months and basically came up with the idea for packet, which was how to connect, what we saw as a software or the world of software with, the world down below, right. The world of intellectual property and Silicon and hardware. And could we do that without getting in the way?
[00:08:18] Could we, can we provide a. A way of automating that experience and delivering it without getting in there and putting our opinion on it. Right. Like a hypervisor and all the other stuff. So call it bare metal now. But at the time we were just like, could we give computers to developers that felt like they actually wanted me to use them.
[00:08:36] Matt: That's interesting because it does, to some extent it, you know, you, you, you know, there's that old joke that all, all, all problems in computer science are solved with layers of infraction. And so the virtual machine and the cloud service does that enabled. You know where the definition of the world at that point, and to many extent, now, and so it's a little counterintuitive to think, well, let's get rid of all that [00:09:00] stuff that enabled the cloud.
[00:09:02] I mean, how, how, but, so, so one thing you said is really interesting. You said we wanted to give something to developers, want to use, and that seems important. So can you talk a little bit more about that?
[00:09:12] Jacob: Well, we just saw, I mean, I use developers as a stand in for software and I love like Andy Jassy last year at AWS, said something to the kin of like, if you're an innovation, then you probably are in the software cause false softwares where innovation is happening. So for software you need developers and developers hate it and love cloud.
[00:09:30] So you probably need to be in the cloud, which of course has a great. You know, narrative right into to AWS. I think it's true. I think that the software is like the macro trend of our world and I'll bet on software any day open source, especially. And that's the problem with software and the opportunity with software growing up with open source in the two thousands.
[00:09:54] And now, especially this decade. There's so much opinion about it, so different. And [00:10:00] so I use developers as a stand in for software. I really think the customer software and. the problem, like I said, it's like, there's so much diversity and there's just more and more. And when you pair that with diversity of hardware, like, you know, you, you talk, I say the word Silicon, but I really just mean like, when anything gets to scale, you start kind of wrapping the, the hardware around the software, like doing it together, like look at Apple, right?
[00:10:23] They, they make every bit of their supply chain work together to do something amazing. And, you know, what we, what we were looking at in 2014, wasn't like, obvious, but it was sort of like if you, if you think on 10 year cycles, which is how long it takes to build infrastructure, you know, because you're doing it right now is, you know, what's it gonna look like in 10 years, way different than the previous 10 years.
[00:10:45] And we thought, okay, well, there's going to be like a lot of things that are scaling up and changing the way we interact with everything probably going to be software-driven and we'll probably care a lot more about the hardware. And so thinking about doing that meant that we would have to design a different one instead of a lot of the [00:11:00] same thing in a few places where like, we probably need to figure out how people can use a lot of different things in a lot of different places, because that's how you create experiences that are unique and personalized.
[00:11:11] And. And I had to kind of run-ins in 2014, that, that really catalyzed, I think what we do today, which was Alex Povey was just starting Coro S and he dropped by New York. And, I guess he had been like backpacking in China or doing something after, you know, finishing with Rackspace and. And we were like, what are you working on?
[00:11:30] And he's like reinventing the Linux, you know, really. And that just reinforced this idea that developers and software were going to work all the way down. So while the cloud was abstracting, like you said, layers of abstraction to solve all the problems that were there, which was repeatability and scalability and whatever software developers were working down the stack.
[00:11:51] And we saw those things in opposition. Okay. And then the other one was actually a friend of mine who had, had invested in them. Zach's previous business [00:12:00] named bill lovey. He actually sold a business to Equinix as well. And you know, I'd said, Hey bill, you know, and you want to invest in packet. And he was like, ah, kind of adding infrastructure, like really what are you into to, and he's like, craft.
[00:12:15] He was investing in all kinds of stuff to do with it, the craft movement, which was like, Hey, what do we care about? Nice. makers, but
[00:12:22] also like, you know, yeah. I mean, what kind of coffee do you drink, Matt? You know, you know, this idea of sort of like, caring about no, but I know the guy who does, so it's all good.
[00:12:36] And that idea, I thought really jived with my sensibility of like who the buyer was and who was creating, was someone who was caring about the craft of their software and the infrastructure and the experiences they're creating. And I remember getting laughed out of the room by my, by Raj Raj who runs for fauna.
[00:12:52] Cause I was like, what do you think about this for a tag line? You know, infrastructure is not a commodity, it's a craft, and he just laughed. He was [00:13:00] like that. So stupid.
[00:13:02] Matt: I mean, that was his motto for years. So
[00:13:05] Jacob: Right. I was staffing and I was just an opera musician. I didn't know Rucker, but, that idea of the buyer being different software, getting way, big infrastructure, deed, more specialized, all, let us to think, how do we get out of the way?
[00:13:18] How do we focus on delivery model? How do we focus on fundamental things? And that's really what we went out to do. And I think it led us to the edge. That's what led us to the idea of the edge.
[00:13:29] Matt: So before we go into D let me put a Thumbtack in this and, for the listeners who may not know what packet it is or why you mentioned Equinix, let me try to describe what packet is and you tell me if I'm right.
[00:13:41]so, when you think of the cloud, you go and let's just pick Amazon after pick a reason.
[00:13:46]you go and you provision an easy two instance. And the reality is that is not a machine. It's not a piece of hardware. It runs on a machine somewhere. but all you're guaranteed is that you're going to get some fraction of some machine somewhere. For the most part. I mean, if you buy
[00:13:59] a large enough [00:14:00] instance, you might actually get a dedicated machine E but you don't ever see it.
[00:14:03] You never actually get to touch it. Like you said, you never get to touch the bespoke hardware. And it's, it's the power of it is they're all the same. And what you're saying is, well, that cloud was huge. It doesn't serve all of the developers that there's a class of developer that cares about the craft, or maybe cares about some other things that they can't do.
[00:14:22] In a virtual instance somewhere, or in a location that you know, is well suited for lots and lots and lots of racks of servers. And that's when people come to pack it. Is that, is that essentially it?
[00:14:35] Jacob: Yeah. I mean, when you have more opinion basically about what it is, where it is, who owns it, you know, it's at the latest and greatest, or is it the one from nine months ago? Do I have access to. The bios or not, you know, all the things that I would say are not generic. So I like to describe packet in the mail infrastructure business.
[00:14:54] We're really, if you think of the cloud is sort of like retail banking, you know, there's like it's good [00:15:00] for 80 or 85% of everyone. the features aren't that different when you go to wealth management, but the. You know that the buyer's looking for something different and it could be. And I think what we see with our customers is that you've got everything.
[00:15:14] You've got stuff that's generic or stuff that's kind of generic or stuff that needs to scale. And then you've got stuff that just is different. And I think we're betting on the fact that there's going to be, A real growth in the, the number of companies, that tried to create amazing experiences to when mobility or when entertainment or when
[00:15:38] Matt: And if I'm trying to squeeze that last 10%, I'm going to want to go to the hardware
[00:15:43] Jacob: You're going to just care about all of it. I mean, I don't think it's always about hardware. I just think it's that you're going to care about the whole thing and the whole thing, obviously it's about cost. It's about performance. It's about security. It's about where it is. It's about so many things. We're not smart enough to know those.
[00:15:57] I think. the challenge is really [00:16:00] that we have a hyperscale world supply chain and distribution. We're good at that. You know, the crowd providers have nailed it, but when you start doing subscale, like you want to do two racks in Detroit. Good luck. Right. And that's the business that I think we're in as we look at, not only edge versus core, you know, that kind of concept, but just as we look at businesses being.
[00:16:24] All digital businesses, they're all going to be digital businesses. They're just not going to any, not everyone there, the restaurant getting to get out, but like the ones who really go and try to lead those spaces are going to be all in on technology.
[00:16:37] And they're going to need a lot of help.
[00:16:39] Matt: Yeah, it's, it's really interesting. the more our businesses become digital, the more we're dependent on like actual mechanical things. you know, I come from the cloud world. And so the fact that data centers have moving parts was actually a surprise to me. I mean, if you think about it, of course they do, but you know, servers have fans,
[00:17:00] [00:17:00] right.
[00:17:01] There are these, these cables that people have to like figure out how to route. Yeah. And they didn't make a lot of noise. Yeah. Yeah. So it's, it's really interesting how much infrastructure and how many people have to care about the infrastructure to actually actually make it work. Let's, let's talk a little bit about, about, about edge.
[00:17:19] You know, we, we throw that term out there. you're using it in a way that, that I think expects that. That someone like me. And because we sit on the coaching to this day, the edge, we probably are talking about the same thing, but just for the audience basis, when you think about it, it's like, what do you think about like what's happening at the edge and where's packet on the edge and how you relate to everything else that's happening on the edge.
[00:17:38] Jacob: Cool. Well, I'll certainly things are changing and, I think it's a good question. Like, what is the edge mean to you versus what does it mean to me? And that's, I'm going to be different. So for us, you know, we've become part of Equinix. So we required, and close the transaction in March to become, you know, part of Equinix and what they were trying to do, was to bring in more.
[00:17:59][00:18:00] automated and operated experience, accessing connectivity, ecosystems. That's the business that aquaponics would, I think describe themselves as being in and connecting businesses and connecting networks and connecting things. And, you know, currently the way you do that at Equinex is you, you know, get a rec, fill it with servers and do it yourself.
[00:18:20] And that's hard for, for, for not only some companies, but to do it at a global scale. I mean, Equifax now has. You know, 210 or so data centers and 60 markets. This is a lot of places to go if you are interested in being that. So when we think about edge, at Equinex, we obviously think about the interconnection edge.
[00:18:39] And I think that that's, you know, a little bit of insider baseball for people who come not from the, co-location or the earlier days of the internet, which is like, how does the internet work again? It just works. I don't know. Right. If you come from a cloud world, it just works. It's kind of amazing.
[00:18:56]but you start getting down there to the builders of the internet, how the internet works. It [00:19:00] works for rap and proximity and. You know, physical connections and the fiber and all that stuff and where those aggregate is a really important center of gravity. And, you know, I think it's not too far to say that for a lot of companies or even for a lot of platforms, the edge is really at those interconnection points.
[00:19:19] That's where they need the telco operator or the, the ISP or the enterprise network and the cloud. And that's. You know, from a physics standpoint, biological edge.
[00:19:29] Matt: Yeah, and I think it, it really makes a lot of sense too, because, that's where the internet. Backbone, so to speak begins and ends. and so once you're on the backbone, you're probably long hauling, to some centralized data center campus or something. Right. but at, at the off ramp of the backbone, so to speak, is an internet connection point.
[00:19:47] And, you know, your parent company, Equifax has, you know, done a masterful job in
[00:19:53] enabling these interconnection points.
[00:19:55] Jacob: It's been the focus, right? It's been their focus since they were started in 1998. And [00:20:00] I think that that's something we really relate to. And like, we obviously came in with kind of a, almost a religious focus on like what we do and what we don't do. Right. And it's like, we don't go up the stack. We go down the stack, very sort of like mission driven.
[00:20:14] And with that clinics, they're really, Very serious about being in the business of interconnection, of connecting things and being in service to that. Now what it means from the edge. And I was going to kind of go there next. I think things are changing as I think that we are seeing, you know, some real early, but really strong examples of like, well, if I could do it differently, maybe I went to it differently.
[00:20:35] Like look at, You know, smart retail, right. Or delivery, you know, kind of, you know, six months ago we would have said that's important and that's emerging than in a post, you know, COVID It's like kind of urgent. I was making fun of it the other day. It's like thinking about, you know, remember that New York times about three years ago started doing virtual reality, like [00:21:00] reporting like a little bit, like, and it was
[00:21:01] Matt: I don't remember that, but yeah. Well, I
[00:21:04] Jacob: with it. Yeah, the cardboard thing. And I was like, Oh, that's cute. But like kind of silly. And now I'm like,
[00:21:11] actually it's really good to do.
[00:21:12] Matt: Yeah. Right, right. I can't get on an airplane, but I'd like to go visit the Eiffel tower
[00:21:16] Jacob: That's right. Yeah. I mean, I'm attending a conference next month with Oculus, you know, parent I'm like super like, okay, well that's like maybe not the way I would prefer to do it or prefer to fly to Cambridge and hang out in a pub and go to do all the things that we as humans want to do. but I can see it kind of reinforces this idea that, you know, technology is, critical to what we're doing next.
[00:21:39] And we have to do it in a more efficient way. We have to do it in a more scalable way. And that means in my view that we have to continue to be agile with the technology, the hardware, the software, the networks that pull it all together. The building blocks that if we use generic building blocks, we're going to make a lot of the same things.
[00:21:57] And I think that those are going to be pushed [00:22:00] in new ways. So I think the edge is going to shift. I think we're. We're kind of, you know, seeing, you know, which workloads and things will pull us there fastest, but otherwise we're going to continue to change that definition, which I think is revolutionary. I mean, we really haven't seen the edge of the internet shift since it got started.
[00:22:19] Matt: Yeah. In fact, in fact, in the, in the most recent, state of the edge report, which shameless plugs stateoftheedge.com free to download, it's an open source project. So I don't feel too bad about plugging
[00:22:28] it.
[00:22:29]but we talk about the, the, the three acts of the internet and, know, the, transition point from the first act to the second act is women, the CDMs.
[00:22:38] you know, at that point it was , and it came out of some research at MIT. In fact, when I did the research, cause we were like, let's get down to the bottom of what edge computing deans, because there's, you know, a hundred different answers. And my first. The first reference to edge computing that I was able to discover in the literature was the research paper that folks embedded, maybe it existed prior to that, but that, that was when I first [00:23:00] edge computing the way you and I are talking about it today, first existed, but that was one way and it was caching.
[00:23:05]and now we're entering a world, this sort of, you know, third act of the internet where. You're, you know, a lot of things are changing. You know, I talk about how, how we're moving from a world of primarily humans talking to machines to primarily machines talking to machines and machines are relentless.
[00:23:21] First of all, they are impatient. Right. You know, we'll write, we'll wait, we'll wait. Ones of seconds to refresh. Facebook will not be happy about it, but we'll do it. but for our machine, you know, that the car just traveled the length of a football field while you're waiting for Facebook to load,
[00:23:35]
[00:23:35]And so there's, it's creating these, these new demands on, on the network.
[00:23:40] And so what I hear you saying from packets perspective that at least today there's a major nexus point, which is this, this internet exchange point where all these networks come together and, and, peer or exchange data. and that's kind of an. An ideal place to at least begin starting to build out some of this hardware [00:24:00] infrastructure.
[00:24:01]how do you view the continuum from, let's say that that, well, let's go all the way, the core. How do you, I mean, so there's the, the core internet and there's the, you know, long haul fiber or the BGP routes, then there's this, this, you know, internet exchange point, the sort of typically on a regional basis, which might be.
[00:24:19] You know, hundreds of thousands of miles, depending on whether it's a tier one or tier three city from, you know, the end user. And then there's all this space between the I X point and, and the last mile network, which is still has to, has to get to the user and the user. And we have a lot of, you know, what people are calling edge computing that sure.
[00:24:38] It looks a lot like on prem. So how do you, how do you deconstruct that whole spectrum and how do you, you see them working together and driving or cannibalizing each other?
[00:24:47] Jacob: Well, I definitely think, that last place working together, I mean, I is like, it has to, right. I mean, in a way that's why Equinex even got started because the internet looked like it was going to break in the nineties because it was like, [00:25:00] nah, we're just going to make it private. We're not going to interconnect.
[00:25:02] We're not going to share, you know, and that's become a huge
[00:25:07] value to the internet as it,
[00:25:09] Matt: Yeah.
[00:25:09] Jacob: scaled. Right. And so I think it will be about working together and that it's not like I'm clear, you know, what, what, what people are applications or whatever value. And so you have to have ways to exchange and deal with that and connect.
[00:25:26] So I think that we will be interconnecting and we will be working together, which I think is at the core of the spirit of the internet. And I hope it thrives. And I think that that's a good idea. What I think we'll see is that first of all, software is going to get way smarter about stuff that we're talking about.
[00:25:43] Like we're talking about like, you know, well, we're talking about things like, you know, hundreds or thousands of miles, right? Well, proximity is a complicated thing when it's thinking about traffic and, congestion and, you know, direct routes or other, that
[00:25:58] means it's just. [00:26:00] just very, and you know, I don't think that we're going to escape sort of like, well, everything's not treated equally.
[00:26:06] Like there will be fast lanes. There will be ways in which, you know, people are like, wow. I happened to launch a few thousand satellites over the last few years. So maybe I'll get to Singapore a different way, you know, for money. I think that there's a lot that's going to change that way and that what software will do is better understand those options and software will help us make.
[00:26:25] Better choices. And so software needs to be invited in. I think this is really where the world of suffers is huge, but where I'm most interested in and evolving right now is learning how to understand networks. You know, we learned a lot about computers with software over the last few years, call it cloud native, but we've just really scratched the surface in a way, because like, like I said earlier in the cloud just kind of works.
[00:26:49] So one of the reasons that it works. Is there, you don't have to worry about all that network stuff.
[00:26:54] Matt: Right. Right. Right. So somebody, somebody lower lower in the stack is retrying. Your packets
[00:26:59] Jacob: Yeah, [00:27:00] exactly. Exactly. And, actually we called our, our company package. So the one thing you have to buy it from a cloud providers network, right. They have a lot of influence over you there, but I think that we're, we're inviting more people in, and especially with the edge, it's a huge opportunity to bring the world of software, call it developers into the world of physical networks that currently power the internet.
[00:27:20] And I think very well.
[00:27:21] So that's a
[00:27:22] really interesting
[00:27:23] opportunity.
[00:27:24] Matt: Yeah, that's really interesting. I think, I think that we've thought of the internet as a bunch of computers with a network that connects them and there is potentially an equally, maybe even more valid way of doing it. It's like, no, it's actually a network with some
[00:27:39] computers
[00:27:40] Jacob: That's right. I mean the inter internet. And I do think, you know, we experience, right now, especially in a very, you know, shelter at home kind of world. We're experiencing a lot of things through. through the lens of connectivity. And so that's a really interesting, like obviously you're going to do something with that.
[00:27:58] And so I think I [00:28:00] would make a judgment about one being more important than the other, but I think that there has been a lack of software innovation, on the network side and just understand the physical world, which is like, you know, I always use the analogy. My brother gave it to me one time, which is like in the U S we have, you know, phone numbers that historically before cell phones were.
[00:28:19] You know, a way in which you could find which central office
[00:28:23] you are. So
[00:28:24] Matt: literally routing codes go, go left here, right there
[00:28:27] Jacob: exactly, exactly in my, my, you know, phone number growing up was so yeah, one, four, nine, seven zero, which got you literally choose the neighborhood central office. And I think that the equivalent of that doesn't get exists for the edge and for being in thousands of places.
[00:28:47]with your application, whatever it is and what to pay for it. Is it all going to be, you know, 32 cents an hour kind of stuff, or is it going to be more dynamic? I think we have [00:29:00] fun.
[00:29:02] Matt: What do you, what are you, what are your predictions? I mean, it's hard. Yeah.
[00:29:05] Jacob: Yeah. My prediction is that it'll look a lot more like, other kinds of dynamic markets where, where inventory is limited, and varies all the time.
[00:29:13] Because if there's one thing that's like really obvious about the edge versus the core is that there's just like a lot less room, right?
[00:29:22] And so
[00:29:24] Matt: Which is
[00:29:24] kind of the universe of the cloud. We tend to think of the cloud is just infinitely, scalable. Like you, you need another instance. You could another instance. Yeah, you're right at the edge. Cause you talked about, you know, the, to the, the, you know, the two racks in Detroit.
[00:29:36] Jacob: Yeah. If you were so lucky, right. You know, and then what if you need certain kinds of things, you know, what, if you need anything special. And so I think
[00:29:44] Matt: Like what's special with a GPU
[00:29:46] Jacob: yeah. What if I, I can only run it where I can. Or what if I need, DPDK offloads? What if I need SGX to secure my,
[00:29:55] you know, enclaves? I don't know, whatever my requirements are says, BMW. [00:30:00] Yeah. Or, or wherever I'm willing to pay for it, you know, there's all kinds of nuance there. So I think, I think that'd be super fun to see, software get super smart about all the things, the assets, including the network and help us create a marketplace. That looks a lot more like, I mean, it's not a great time to be in the airline business, but something that hotel and the airline business did over the last 10 years was figured out how to create
[00:30:22] amazingly
[00:30:22] dynamic pricing, you know?
[00:30:24] Yeah. And I think we'll move away.
[00:30:28] Matt: points? Right. You run a bunch of workloads late at night when nobody else is doing it. We'll give you a free one in the morning.
[00:30:38] I
[00:30:38] Jacob: think this is,
[00:30:39] this is how
[00:30:39] it will go
[00:30:40] Matt: yeah. And probably the biggest, yeah. France is that, these decisions will need to be made thousands of times per second,
[00:30:48] Jacob: all the time. So I think we need, we also need clean. It's tougher to get smarter. and that's, that's tricky because we have to open up things that are traditionally not very open, you know, things like private network. So [00:31:00] not
[00:31:00] Matt: Well, yeah, I was going to ask you about that. So, so it, it, it, you seem to have a strong opinion. one that I think I agree with. Which is that, in order for the software to get smarter, it has to get much better at understanding like what's there, what's running on what, what the routes are. you know, what, what the, what the reliability is, what the network congestion is.
[00:31:19] So
[00:31:20] how, how is all that information getting surfaced?
[00:31:24] Jacob: Yeah. Well, I think some of it's our job is to create standards with each other in as service providers to do the equivalent of seven one four nine, seven. Oh. You know, and, and figure that out and, you know, standards then create, I mean, I, I'm not, I I'll say it again. I'll bet on software. Anytime I think properly served up access, innovation will occur.
[00:31:45] And that there are a lot of, innovation minded people touching, internet infrastructure. And so it's kind of in our best interest, instead of guessing the use cases and saying, I know what it's going to be, I know what it's going to be. The one that's going to make [00:32:00] all the money. I'm the one that's going to make it all make sense.
[00:32:02] Right? I need to invest in these super expensive longterm things. When the move in the world is moving so fast, I think it's better to the other way and look at how to be more open. Like you look at what of the things the clouds have done super well. They've created ecosystems. They've created ecosystems where.
[00:32:17] If you're in the marketplace, you're going to do more business. And so there's like a lot of reason why you should invest in that. And I think we can create the same mindset around the inverse, which is this diverse, totally wild West cleared problem. Special maybe only matters to them a few or a few hundred companies, world called the edge.
[00:32:41] Matt: So, so do you see a world where, cause you mentioned, you mentioned interoperability, in standards. so do you see a world where a software program can say, I need a computer that looks like X, like a declarative statement and kind of put it out there and you know, not just a packet to maybe, you know, packet prime. [00:33:00]
[00:33:00] and, and someone tells me, yeah, if you, if you're willing to pay a nickel, I'll run that workload. I mean, is that, is that the future?
[00:33:07] Jacob: I think it's, I think it's definitely a part of it. I would love to see more of that. I think, I mean, I've never been a blockchain guy,
[00:33:13] despite my
[00:33:14] pokey net, maybe,
[00:33:15] Matt: that a blockchain company
[00:33:17] would go,
[00:33:17] Jacob: But I stepped away from it.
[00:33:19] Maybe it was a good idea. but you know, I've never been sort of into the crypto, you know, kind of scene, but I'm really intrigued by a lot of the work happening there around, highly distributed
[00:33:28] computer and change the trust. And I think that there's a lot of value in that, you know, in the, the core sitting on our phone that we're not using most of the time. And this is where we as a world, we have limited resources. Alright, and we will need to become more efficient as well. You scale these things up and become more, you know, driven and billions of people are using them.
[00:33:52] And I think that's, that's kind of the answer is, is inviting that kind of innovation and, and utilization. And [00:34:00] so interoperative abilities is key. And I think software is kind of built to be that way. We see it with cloud native kind of changing the world in four years. It's
[00:34:07] not because it's the best tech Everett's because it allows you to be portable,
[00:34:10] Matt: Yeah, you said something interesting and I, I, I'm not going to get it right, but I'll get, I'll get a close, I'll paraphrase it close, but it was essentially, you know, access served up properly, you know, but gets innovation. tell me about that. Like where, where have you seen that in the past and where are you seeing it now
[00:34:31] Jacob: Well, I mean, I think access is key for learning for tooling, for trying for failing. Right? So for me, some of that access was at 24 hour computer lab, you know, at Carnegie Mellon or the fact that someone was like, Hey, there's an MTC that the orchestra, if you are willing to suck at the bassoon and come, come on down, right.
[00:34:49] I think just the ability to go and do and learn is important. From an access standpoint, like we've had a lot of, initiatives we partner very closely with the world of still could because we don't [00:35:00] by definition abstract people from it. Right? So you want to buy, you know, arm based processor and until this processor and AMD based processor from us sounds good.
[00:35:09] You know, you want to be very opinionated about that, where we welcome that. And so we've had some neat partnerships and I've been amazed. We did one a couple of years ago when arm was first starting to try to get in the data center space and, You know, like we, we had gotten some investment from soft banks and we were looking at this a lot and we called up the guys at armor.
[00:35:26] Like, what if you just gave it away? Like, you know, through like an API, like to like the right people. And they're like, really, actually it would be really helpful for us. And so we came up with a program called works on Arma. We just gave infrastructure to open source so that they could build some buffer on it.
[00:35:43] So that then it would be upstream naturally. And, It's been a huge success or, I mean, I think part of the success of, of arm kind of going from a gotcha. It just doesn't work for me. Cause my software doesn't work on it to being like we're pretty good. And we saw the same thing with, the [00:36:00] CNCF where we donated infrastructure there.
[00:36:02] And really, it was just like completely unopinionated and we, the weirdest things are attracted to that. So I think lowering the barriers, you know, for access are important when we want to see innovation, that's not obvious. But maybe it takes a little bit of leap and that's where I think we can collaborate.
[00:36:18] You know, that's where collaboration is valuable to all of us, even if we're competing with each other. Right. Because the bigger,
[00:36:30] Matt: Yeah, and it's and the power of a, of a platform. And I know that that term is overused, but, but I, I think the spirit of it, we can probably agree on, the ability to, you know, whether it's right once deploy many, or to, you know, utilize tools that remove the complexity, whether it's because you've got fungibility developers, meaning, you know, everybody that, that builds apps for one of the major clouds has a job somewhere.
[00:37:00] [00:37:00] Right. And so you create these, these, these, and open source. So you mentioned open source earlier and open source is a platform itself. you know, you think about like an example that I love to use is the 35 millimeter film market, which is, you know, something that predates most, most people today. But you know, back in the day when, like they don't film, you used
[00:37:20] actual motion that reacted to light and physical shutters and all that stuff.
[00:37:25]But the camera manufacturers, the film man crush got together and they say, look, let's compete on the quality of the emulsion, the, the, the, you know, the, the, the sophistication of the lenses, the speed of the shutter mechanisms, the weight of the camera, the features, but let's not. Compete on the platform, which is the, how the distance, the Sprockets,
[00:37:45]
[00:37:45] in the film.
[00:37:47]and I think to some extent, you know, opensource is like that. It's like, how can we all invest in this platform that raises the tide for all shifts that we can compete
[00:37:58] at this higher level, which
[00:37:59] is on [00:38:00] what we build on, on top of. Yeah, the top of the platform. and you know, and I actually see that analogy with state of the edge as well.
[00:38:07] And talk a little bit about, because, you know, the origin story exists. We're not in this room because we're actually doing this over zoom, but the origin story exists in this room. And the fact that it's now an open source project, sort of interesting. So, but why don't we transition to the state of the edge, which for those people that don't know, it's a project at the Linux foundation, it's one of the top level founding projects of LFM edge.
[00:38:29]it's pathway into Linux foundation was, over time and interesting. It started with the open glossary of edge computing. but let's, let's roll back before LFS even existed before we had anything.
[00:38:39] Jacob: Well, I remember it was like at the end end of one of these, you know, kind of whirlwind, it was probably a VC pitch trip where I was like, we're running out of money. We need to really.
[00:38:54] Go to San Francisco
[00:38:55] Matt: you're you, you want to buy, you want to build infrastructure,
[00:38:58] Jacob: I know. Yes. [00:39:00] Many knows that only one. Yes. Not so much. and so I probably really needed that beer that we had, but I remember we got together cause we were switching out people who were just doing stuff that, you know, as my friend, Andy Schwab, Becker from SoftBank would say, if you loved 4g, wait until six G, which is like the feature.
[00:39:17] Is never really clear. There's just people working on it and then find out later. And we were looking for people like a great grid raster and, you know, you know, various people who were working on new, different, totally impractical things. Right. And, I think that's what brought us into the same room frequently, including.
[00:39:38] That outdoor beer garden to kind of like we should do. And it was one of those crazy ideas sessions. Right. which was, I think crazy in a way that's very grounded in being collaborative humans, which is like, we should work together on something. And I love that kind of feeling that, that open source
[00:40:01] [00:40:00] Matt: Yeah. You know, I never, I never thought about that really, but it's, it's, it's very much the case that working on the state of the edge. I mean, you know, there, there, there were enough participants in, in certainly this last year that some of us competed with each other.
[00:40:17] And yet when we were working on state of the edge, we took that hat off.
[00:40:21] Jacob: Yeah.
[00:40:22] Matt: creative project. That's a really interesting point. I
[00:40:24] Jacob: And I don't think it's just altruism. I think there's some of that because I feels good. And I think for some of us who've benefited from. From things like open source or other kinds of collaboration, you know, I grew up playing in orchestras where maybe some people play sports, you know, where you come together as a team.
[00:40:41] Right. And sometimes you're on the same team. It's sometimes you're also just doing a pickup game and that that's a really human. Thing, which we miss now that we're not able to be together. Right. But the idea of, collaborating, not just for altruism, but because it's in rational, self-interest, isn't on Rand would say, right.
[00:40:58] It's like, it's good for business. [00:41:00] Makes sense. I think that that's a really strong reason behind the state of the edge and behind other things like that, which are like standard, you know, like let's figure out that problem in the film and you know, this is just logical. It doesn't always work out.
[00:41:14] And sometimes I think say that the edge came at the right time, frankly, like we, we came to the conclusion that we should collaborate and invite people in to help get a common language around the edge and try to give it a voice that was neutral. because there were other things floating around that were, you know, this company talks about their edge stuff and this company talks about that, but we went in different direction.
[00:41:33] I think that has really helped to set the tone for the edge. Like we are going to work together.
[00:41:45] Matt: Okay. So, so, so when it comes to edge computing, cause this, you know, presumably is a podcast about edge. What do you, what is, what has changed the most in your mind?
[00:41:56]Jacob: well, I'll tell you what I think hasn't changed, but I think the edge actually hasn't changed much. [00:42:00] I think the potential of it is much clearer. Yeah. Well, I think we're still getting gravitated back to like opened up, you know, some sites with you guys in Chicago. And of course we leave back called it the back, held a lot of our traffic over to the clinics in Chicago.
[00:42:13] Right. I think that's still a thing. Right. And that's okay. But what's, what's really changed completely. I mean, I just think like it's easy to forget how. Like portability of like, not just workload, but like all the workload, like databases and weird stuff and everything three years ago, that was like, I mean, there were people hoping for it, but it wasn't really happening in a, in a scalable way.
[00:42:39] And now we have like so much progress on the ability for frankly, I'd say fairly average. Technology ina users to deploy stuff in a lot of places. And I think that's a huge shift that's going to enable so much. cause like you said, you were talking about CDNs earlier. I mean, [00:43:00] how many, how many really, you know, how many companies really had infrastructure at a hundred or 200, 300 places?
[00:43:08] Like not very many
[00:43:10] like 20.
[00:43:13] Yeah, maybe. Yeah. I mean, if you start looking at a couple of tacos and whatnot, I think now like three years ago, that would have still been the case. You would have been like, raise your hand if you have infrastructure only in us' and that would have been most people like, Oh, are you in lots of locations called USC and us West,
[00:43:29] but now I think you're seeing the potential and even the reality where you could find people being like, yeah, absolutely.
[00:43:35] I mean, I could deploy this. Application or this part of my application globally. And I think that's a huge mindset shift and that's like wind behind our sales here as we kind of figure out the next chapter.
[00:43:47] Matt: Yeah. And I, you know, I think of, you know, it's sort of interesting, the tussles we had in the industry, back in 2010 and 17 are very different than the tussles we have today. I mean, back then, one of the biggest arguments is [00:44:00] weather. Your edge workload, and I'm oversimplifying, this is going to run on or next to the device out in the field, or whether it's going to run on or next to the infrastructure, you know, at an X point or at the base of the cell tower or at a, you know, some other, some other, you know, fiber nexus or whatever.
[00:44:19] And it turns out the answer is yes. Right,
[00:44:23] right. Yeah. And I don't think, I don't think anybody really ever thought. The airbag of the car is going to be deployed over a five G network by some cloud workload that is subject to, you know, a bunch of different variables that a tight, you know,
[00:44:42] Jacob: I think as we don't know, but yeah, and I love, I mentioned my friend, Andy SWAT, Becker, is that he has that SoftBank and. And it was an investor in a sense, he gave a presentation one time at our conference side effects that are so great. It was like, you know, if you loved four G wait until six G, [00:45:00] which it was, it was really funny because you think about 4g, which is the most recent evolution that we have that we can all remember in watched happen.
[00:45:08] And like, what was the big killer app the iPhone was created and created the app store? Like there wasn't,
[00:45:16] Matt: The eliminate the elimination of the voice call.
[00:45:18] Jacob: Right.
[00:45:19] Matt: That was the big innovation.
[00:45:21] Jacob: and then, you know, Steve jobs shows up with this like magical device or whatever they call it. And, for G was, was useful for that. And, and suddenly then you had a reason for 4g, but you didn't have it before.
[00:45:33] And we had to build it all first.
[00:45:39] Matt: Yeah. You know, you know, it's really interesting that you talk about, you know, building platforms. Can you talk about, you know, layers of abstraction and tooling that, that brings some commonalities to solve some of these harder problems? and the iPhone is actually a really good example because, you know, one of the thought lines that you introduced earlier was this idea that we actually don't know what the killer apps going to be.
[00:45:59] I mean, [00:46:00] we have some hypotheses. I mean, you have a sense of where those Venn diagrams of things that people will find valuable, things that are now capable with this new infrastructure, the software tooling it's available. There's a fin there's, you know, there's overlapping sense there. And we have, we have this intuition that if we build.
[00:46:17] Something, these imaginative things you think about the iPhone, right? Like, I mean, when it was launched, it was launched as, okay, this is a phone you can make voice calls. It's your, has your music collection. It's an iPod and you can do email right. Or browse the web. Like that was, that was right. That's what it was.
[00:46:32] And not that much farther out came something that we all take for granted. but at the time Apple, the miraculous and that was Uber. And the only reason Uber existed is because. Someone was crazy enough to ship a phone that had a high resolution display, you know, dependent on, on your satellites, in the air for GPS that somebody else put up and at LT network that somebody else put up and all of these things.
[00:46:57] And suddenly you've got the ability to, you know, call a [00:47:00] car to go wherever you want. And transportation arguably has been completely reinvented because of that.
[00:47:05] Jacob: Yeah. Well, I think it's, it's just a truism that, you know, we don't know the, I mean, we can all take bets and I think that that's an exciting place to be is taking big bets and trying to be right and figure it out. but it's, it's really exciting to see the velocity with which, I call it the edge is, is moving where it's just like, whatever's next is definitely coming.
[00:47:27] You can feel it in the air, everything like we talked about earlier, like VR through the New York times or that drone delivery, that seems like literally crazy a year ago. You're like, I mean, really like, I mean, Okay, well, maybe it's not an, exactly what we thought, but like the idea of it is no longer impossible.
[00:47:45] And so I think that's where access and access at the edge and access to information and having a shared language, all the things that the state of the ed is trying to start, how, you know, our community with, I think are really important foundations, but where it happens next
[00:47:59] is going to be [00:48:00] very exciting to see.
[00:48:01] Matt: That's that's amazing, Jacob. So we're at the top of the hour. I guess I'm going to ask you one more question, because I don't know that I have a good answer for this, and I'm curious what you're going to say, and I'll see if I can catch you off guard. if you could wave a magic wand and change one thing in the edge industry today, what would it be? What's the one thing you'd sort of push over over the edge, so to speak,
[00:48:21] Jacob: Well, the thing that I would, and this is totally self selfish. but I would really love to see, see us figure out the physical parts of it. It's like, I mean, some of us are passionate about that. Like the, the logistics and the delivery and getting the things in and moving stuff from Chicago to Dallas and blah, blah, blah, blah, blah.
[00:48:41] Right. But most people don't care. And I think like standardizing or figuring out a way in which we can stamp out this thing. The 35 millimeter thing and really compete and bring value at a different layer would be huge for the edge. And so that I think we could take further and that would [00:49:00] require, I think probably some pretty big collaborations to happen.
[00:49:03] Matt: Well, I'm going to have you back on the show in a year and
[00:49:05] we'll, we'll see how it is. Well, thank you so much, Jacob, for spending an hour with us this morning,
[00:49:12] Jacob: It's not a cabin now. Absolutely. But I'm just amazed the copper wires, road worked for, for this amount of time. So it was great spending it with you.
[00:49:24] Matt: Yeah, it was. Thanks a lot, Jacob. And appreciate you, you know, showing up.
[00:49:27] Jacob: All right. See you soon, man.
[00:49:28]