Over The Edge

Open Source Collaboration is the Only Way to Scale with Jason Shepherd, VP of Ecosystem at ZEDEDA

Episode Summary

Today’s episode features an interview between Matt Trifiro and Jason Shepherd, VP of Ecosystem at ZEDEDA. In this interview, Jason discusses the significance of open source collaboration and interconnected ecosystems in scaling IoT and edge adoption.

Episode Notes

Today’s episode features an interview between Matt Trifiro and Jason Shepherd, VP of Ecosystem at ZEDEDA, an edge virtualization company offering solutions for IoT Edge orchestration. 

Jason left his role as CTO of IoT and Edge Computing at Dell Technologies last year to join ZEDEDA, with the stated goal of bringing IoT out of its “AOL stage.”

In this interview, Jason discusses the significance of open source collaboration and interconnected ecosystems in scaling IoT and edge adoption.

Key Quotes

“Open source is the new way to drive standards and standards drive scale. If you don't have some sort of open source model going forward, it's going to be difficult, because ultimately it’s about scaling interconnected ecosystems, and that does not work without open.”

“It's about ecosystems. What you want is an ecosystem of lots of participants that all have a shared interest in building a platform upon which they can differentiate. So build one or join one, but either way be a part of one or you're going to struggle going forward.”

“Open source in general has just become the modern way to drive standards. Collaboration and open source– that shared technology investment– it minimizes undifferentiated heavy lifting. Stop doing that and start focusing on value. And it helps you create a snowball effect for standardization. What open source is creating is de facto standards. If you get enough people using something, it becomes a de facto standard. It's just a new way of driving that network effect for standardization

“The next domino that you have to solve is trust. You've got to figure out a way to scale trust over heterogeneous networks or none of this stuff will ever work. It’s about building trust at scale, and using technology to help you.” “Open source is the new way to drive standards and standards drive scale. Do you think Google didn't make money off Android and all the ad revenue that that ecosystem drove?”

“You'll never scale to the grail if you don't take an open path. Even if right now it's about solving  simple problems and creating some value, you'll never get to the ultimate goal of entirely new business transformation across many markets if you don’t have an open base.”


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 Ori Industries. Ori Industries is building the world’s largest edge cloud. Their products power the next generation of intelligent applications through unparalleled access to major communication networks worldwide. Ori is laying the foundations for application developers to seamlessly deploy to uncharted edge computing infrastructure across the globe. Learn more at ori.co


Connect with Matt on LinkedIn

Follow Jason on Twitter

LF Edge taxonomy white paper

Project EVE at LF Edge

EVE in the Market

Project Alvarium

Episode Transcription

[00:00:00] Matt: Hi everybody I'm Matt Trifiro, I'm the CMO of vapor IO and also the co-chair of the state of the edge project of the Linux foundation. And I'm here today with Jason Shepherd, the VP of ecosystem from ZEDEDA. How are you doing today, Jason?

[00:05:36] Jason: I am great. Thanks for having me.

[00:05:38] Matt: Yeah. Terrific.

[00:05:39] So before we get into all kinds of interesting things, edge and. IOT and Linux foundation and ZEDEDA. I just like to start back at the beginning. I mean, how did how'd you get into technology?

[00:05:54] Jason: Yeah was cool. If we start at the beginning first, the earth cooled, then the dinosaurs came and I'm like, no.  no, but  I I've always been interested in tech. Like I made my parents crazy, pressing every button I could find, and you've gotten to thinking things and building cars and it's just all kinds of stuff.

[00:06:11] And,

[00:06:12] Matt: better at taking things apart.

[00:06:13] Jason: Yeah, muffin and met together.  I just always been interested in, in, in, in technology, but also, but more so for how it like impacts people for good, not necessarily like all the, you know, not some bolts in all cases. Although I am technical, I would, I would always say like, if it's fuzzy, I'm on it.

[00:06:31] So I've of been fortunate to get onto a lot of different tech trends. I started actually as a mechanical engineer following in my dad's footsteps, but then. As I, as I kind of progressed through things, I decided purposely to get more into the software side and really combine that with the business side

[00:06:47] Matt: What kind of mechanical engineering were you doing?

[00:06:49]Jason: I actually had, my first job was at Dell, you know, where I basically designed the, the, you know, with the teams, the sheet metal and plastic, and for the PC and did that for a number of years. And then I went into the startup [00:07:00] world, company called Clare cube. Actually, initially it was a company called,  A wave flight that was doing really cool stuff a lot ahead of its time.

[00:07:06] It was yeah. MP three player that you would stream audio from your computer, where you downloaded a bunch of stuff from Napster to your home stereo. So kind of like a homesteader experience, but for MP3s at the worst time to be doing anything related to MP3 is when Napster got shut down. And so a really fun experience, but short lived and then, went to a company called clear keeping Austin, where I just started wearing a bunch of different hats as you do in a startup world.

[00:07:28] And I went back to Dell and I was about 13 years at Dell and just really enjoyed working there. Money was always kind of on the leading edge of new tech trends, various roles in CTL and R and D and whatnot. you know, that's when I kind of transitioned and in 2014, You know, we're like, Hey, what do we want to do with this fuzzy, you know, IOT thing.

[00:07:47] And so blank sheet of paper from scratch and built up, you know, the, the, strategy with the team, this was pre EMC acquisition. So all of a sudden that little acquisition happens and it just grew. And I just proactively with the team, started [00:08:00] building up our partner ecosystem, got edge X Foundry started in July, 2015, which started as an idea internally.

[00:08:06] And. That just hit 5 million downloads out, open source as our sister project to both state of Vietnam, as well as Eve you know, the data contributed and just, you know, the whole point, you know, long story short is, is it's fill gaps, you know, go be curious, fill gaps. you know, don't step on people along the way.

[00:08:24] That's not cool, but just nobody. I always tell my teams the best way to get a job as are going to be doing it. And so just been able to kind of more over the years and, and, you know, IOT. And then of course now it's fashionable to call it edge, but there's a lot of edges and of course that's why we're working together.

[00:08:39] And why, why you guys are doing important work with the clustery and, you know, it's just continuously improving and surround yourself with good people. You know, that that also help you improve.

[00:08:47] Matt: Yeah, that's awesome. And, you know, you mentioned, you mentioned IOT, which, you know, I'm trying to think of the year that I first, first came on my radar. It was probably 2014, 2015.

[00:08:59] Jason: That [00:09:00] was, that was when it really started kind of picking up. I mean, obviously I have T was going and what 99 internet of things. But at that time we had a lot of customers starting to come to us and you're going, this was at Dell and they're like, I want to buy some IOT. Really? What color would you like?

[00:09:14] You know, it was just like Zift is more of a concept than a yeah. thing, per se, even though it involves quote unquote things and. And, you know, ultimately, you know, when something is, when people talk more about the technology and then the actual use cases, you know, so 2014, 2015, it was like, Hey, buy my IOT platform.

[00:09:32] And it was like, the joke then was 150 IOT platforms. And now then it turns into 450 as the running joke, but it's like, what should I do with it? It doesn't matter just by my platform, you know? And then you can figure out what to do with it. Then in 18 or 19, you started hearing people talk about outcomes and let's talk about use cases and flip it the other way around.

[00:09:50] And, you know, ultimately everything I've been doing with lots of really good people in the industry, you know, YLF edge and all that is we need to drive standardization to scale this [00:10:00] thing. And IOT it's it's as much a business question as it is. You have technology, is there reason to network things and it had, you know, risk and complexity to my life to get value and open, always wins in the end for scale.

[00:10:13] You know what? Dell is a great company. We ship a PC every second of the day. If it costs thousands of dollars to connect your keyboard. No, you know, you must, you know, you need to democratize the South to monetize the North, so to speak South being southbound connectivity, you know, in the, in the it sense.

[00:10:28] And so, yeah, w we got involved with the adjunct stuff, of course. And then, well, if I just, why I chose to come in as a data, you know, again, that was great company, but it was just kinda like. Yeah, time for change. When we get back into the startup world and you know, it just, it just felt like if the right thing to do, I knew the guys when it got to the point where I could replace Dell on my blogs with Devita, I'm like, well, maybe I should just come work for you guys.

[00:10:47] Cause we're very philosophical aligned around the importance of open and, you know, imagine if one company on the internet, maybe it wouldn't have worked out so well.

[00:10:55] Matt: Sometimes it feels like three companies on the

[00:10:57] Jason: yeah, well, well true, but you know, the [00:11:00] ultimate goal here. Is interconnected ecosystems more, you know, IOT is sort of a misnomer internet of thing.

[00:11:06] It's a series of increasingly larger intranet that as you find business value and you can manage IP protection and privacy and all that, you're going to find entirely new ways of things. Coming down. I like to say we're in the AOL stage of IOT right now. Just kind of getting things online, connect that

[00:11:23] Matt: we're getting plastic discs in the

[00:11:24] Jason: Yeah. And the mail and whatnot, but at the same time, you know, and it's not even a technology problem right now, initially. I mean, there's. There's fragmentation, of course, but the problems I've seen over the years, I've been doing this for five years and he said, number one, it's a business problem. You know, is there a reason to do it?

[00:11:42] There's a lot of solutions looking for problems out there. Number two, it's a problem with people. You know, people are your greatest asset, you know, out there and in comes data. But, you know, the old it versus OT equation, you know, line of business often drives things. you know, OT, operational technology, been running processes in the physical world for a long time.

[00:12:00] [00:11:59] Completely isolated from networks because it's, you know, things go boom, and the OT world, if you, if you have a breach, but meanwhile, you have to connect that stuff to get, you know, visibility and drive some analytics and drive change. and it's not about who owns stuff. It's about what are the capabilities.

[00:12:14] Meanwhile, even with an OT, like the production person would compete with the quality person because to increase quality, sometimes I'm pre reduce the stupid safety person, you know, and then even in the IP world and the PC people don't necessarily go on to the server people, Oh, you're not going to put your VDI PCs in my server room.

[00:12:32] You know, we see, we see it across the board. And, and so people is one of the challenges because also IOT data from the physical world exposes. Dirty laundry. Oh yeah. I didn't really realize that I was that efficient. And then case study always become hard. He's like, would you want to share that you were terribly inefficient?

[00:12:51] You didn't know that pump was running down in the basement, you know, 24 seven. anyway, so just there's there's that then you get into fragmentation, then you get into security as a challenge. [00:13:00] You know, if you don't solve problems, one through three business case people. Technology fragmentation, which of course we're dressing and alphaJET and other great consorted efforts out there.

[00:13:10] If you don't have solve problems with your feet, you don't have a security problems. You're not doing anything.

[00:13:13] Matt: Right. Nothing was breaking into, so, I mean, that's one of things I like about you. Jason is everything's so high bandwidth. I think I've had some of the most compressed five minute conversations with anybody.

[00:13:26] Well, let's,  let's, let's, there's a lot to unpack. There's a lot to unpack there. So I'm gonna, I'm gonna, I'm going to pick up a few threads to, to talk out, you know, partly cause our, you know, you and I have been in this in a long time and we can talk a lot inside baseball, but I think many of our listeners are broader, so, but let's start at the highest level.

[00:13:41] So internet of things, the, the definition I remember, you know, probably came from somebody like, somebody out of a wired magazine where, it's like, look, you know, we just, we just did. We just. We're starting to adopt IPP six, although, you know, it still hasn't really been adopted, but there's enough IP addresses to get every device on the planet [00:14:00] and all the ones we imagined coming and IP address, and everything's gonna be hooked up to the internet.

[00:14:05]and I still think that at the high level, that sort of holds, but, I think there's a couple of things that are, that are, that I'd like to tease out. So the first thing is it's called the internet of things, not the. On premise net of things. And so I'd like you to talk a little bit about how all these devices relate to on premises equipment, but also how they relate to the internet and why it's the internet of things and why that's so important.

[00:14:32] Jason: Yeah. I mean, so obviously at the, in the purest sense, you know, IOT and internet of things, and it's about, you know, kind of chronicling the physical world and, and there's kind of two, and that, that can be a certainly on prem, within a building, within someone in somebody's home, you know, a refinery out in the oil rig of mine, you know, out in the boonies, you know, whatever, or of course out.

[00:14:53] Yeah on a vehicle, like some something kind of moving, so either fixed or moving and I'm documenting the physical world and, [00:15:00] and it tends to be upload centric, which is different than, you know, of course down download centric, you know, to client advices. You know, if you're doing cloud gaming, you're care about latency and who's a lot of stuff where you care about latency going up, but generally speaking, it's a different.

[00:15:13] Direction, you know, you're taking data from the physical world and you're uploading it. And then the internet of things does, you know, I mentioned before, it's kind of this notion of interconnecting devices getting that data from the physical world. Initially it starts with monitors, you know, rarely do, do people go straight to analytics, it's just visibility.

[00:15:30] And then yeah, we started applying tools like AI and ML. Like I just kind of assume it's all related. I don't even really like to talk about. You know, IOT is like, you know, the, the end game of sporty AI or whatever there it's, it's what outcome do you want? Then you apply the right technologies. But the, the notion of IOT is kind of like a new enabler for new data, new views into the physical world.

[00:15:53] And then of course, more data that you're, you can analyze, I think is important. there tends to be too. Workload themes that you [00:16:00] see and there there's different, adoption patterns for each. So, you know, I would call it machine intelligence is like, kind of going to, you know, earlier on, just, you know, any kind of structured telemetry.

[00:16:09] So this is sensors in a physical world that speak a certain protocol and there's thousands of protocols. This is why X Foundry, fledge other projects within  matter, it's trying to create more of a universal, you know, Secret decoder ring. and then you've got computer vision, which is, you know, cameras as the sensor.

[00:16:27]and you can tell a lot by the world with a camera, both

[00:16:30] existing.

[00:16:31] Matt: expect things outside of the visible spectrum, infrared and ultraviolet

[00:16:34] Jason: Yeah. People are using cameras for thermal on the COVID cover it. And a lot of people don't cover the response with AI models using computer vision. Computer vision is the killer app for AGI. So nobody, unless you sell connectivity over a wide area, do you think it's a good idea to send four K video over the internet?

[00:16:51] Now you want to analyze it locally and then trigger events. But anyway, it just gives you the visibility. But while a camera can tell. thermal data, you can, you know, the high speed camera can [00:17:00] tell vibration of a wobbling shaft or a motor. last I checked, they can't tell voltage, you know, you need a sensor for that.

[00:17:06] And then even vibration. It's good to have the sensors last. I checked you

[00:17:10] Matt: aim a camera at the volt meter. Read out the

[00:17:12] Jason: Yeah. There you go. Well, people are doing that actually in, in for water meters, like a legacy know water meter. People are snapping a camera on top of it out in the city. They're there, they retrofit rather than breaking the pipe and putting a digital meter in there, like a sensor they're literally putting in an OCR sensor now optic work condition on top of the meter.

[00:17:29] Yeah, we see that all the time. But last I checked too. As you probably even could use a camera as a motion sensor, you probably don't want that as a motion sensor during your bathroom. You know, so it's a matter of privacy and, you know, you know, just because you

[00:17:41] Matt: or the TSA lines.

[00:17:44] Jason: Yeah. So it's, you know, just the whole notion of the IOT it's networking systems, getting that view, the physical world, you know, starting with basic visibility and then advanced classes, analytics, actually advanced class advanced class advanced, advanced to a one and five Oh one to five Oh one is like the analytics of the [00:18:00] analytics.

[00:18:00] I literally look at what's the best place to run that workload to get the optimal results of performance versus cost.

[00:18:08] Matt: Yeah. So let's, let's, let's talk about that a little. Cause I think this is a that's that's a really great, great segue into some of the work that you and I recently did that. Gather with the Linux foundation. So, you know, there's been this funny line that I've used. You ask a hundred people where the edge is and you get 122 answers.

[00:18:24] Right. and, and that's something that, that is true. you know, certainly over the last year, three years, we've seen a lot of convergence on agreement and the Linux foundation across all the  projects came to an agreement and published a taxonomy. And so what I'd like to do, well, first of all, I think that, that when you think about connected devices and you think about them, eventually having a path back to the internet, because for the most part, I mean, there's gotta be, you know, oil derricks in the middle of the ocean that maybe we won't do

[00:18:51] well batch store and forward data collection.

[00:18:54] We're

[00:18:54] not looking at a line of

[00:18:55] Jason: yeah, it's more like a batch. And then, you know, sometimes it's over satellite is you want to minimize, you've got a straw. So you [00:19:00] want to minimize what you're sending over

[00:19:01] Matt: Yeah. Yeah, but I mean, clearly we're moving into a world where, where that last mile communication is becoming increasingly more ubiquitous and becoming increasingly less cost for more data and more devices. You know, whether it's it's more fiber, more 5g or alternative technologies. And so to some extent, the entirety of the internet.

[00:19:20] Everything that we do can be traced from the device or the sensor. That's the smallest thing, all the way back up to some centralized cloud, that might be, you know, in Seattle with Amazon or Microsoft. And there's a bunch of, waystations along the way, the regional data centers, the access edge, the on premise position, things like that.

[00:19:39] And when you're looking at building an application across the totality of that, That spectrum. you have to make a lot of decisions. It's about where you run your workloads. Some places might have more power and be more easily scalable, more easily upgradable. Other places might be faster, more real time.

[00:19:56] And so what I would love for you to do is so walk our [00:20:00] audience through first, the spectrum from, from the smallest device to the centralized cloud and all the stopping points and maybe touch upon, you know, what you should be thinking about workloads that have to run there, or you might want to run there.

[00:20:12] Versus other workloads that you might want to run somewhere else in concert to deliver an end to end solution.

[00:20:18] Jason: Yeah, no, that's cool. Yeah. So definitely send it. Obviously we worked quite a bit on this, this taxonomy paper. So kind of how I'll outline it is, is basically, you know, how we outlined it in the community with an LFS, this taxonomy paper that was just released recently, you can find it online, you know, LFH taxonomy, white paper, but.

[00:20:37] You know, and there's a lot of, kind of different definitions out there, as you said. And a lot of folks talk about it as like, you know, we've got the mirror edge and the far edge and the thick edge and the thin edge and, and the industrial edge and the enterprise edge and the consumer items like, do you know what?

[00:20:50] They're actually not all different. It's one continuum that you apply different tools and considerations to, and domain knowledge to, but, but we need one cake with [00:21:00] pasta flavors of icing and sprinkles and. And it's a continuum and you you're gonna have it's based on an inherent trade offs. And so if you focus on inherent technical trade offs of why you would do something in one spot, or you have two in one spot versus another, you're always right.

[00:21:15] If you use loaded ambiguous terms, you confuse people and you end up with 122 definitions.

[00:21:19] Matt: Yeah. So, so

[00:21:20] what are some of those, the most important trade offs that you need to look at?

[00:21:23] Jason: Yeah. So number one, you know, it's, it's, am I on a wide area network relative to the user or a process I'm serving where I'm on a, on all levels Clarion network and the reason why that, that matters and maybe another way to put it in other ways is that latency critical.

[00:21:36] And this is in your glossary, you know, within the edge, of course, is that latency critical, meaning. I have a safety issue or I lose a hundred thousand dollars. You know, if, you know, in my process, if something goes down or is it latency sensitive, latency sensitive, you know, like Netflix, no one's going to die.

[00:21:52] If my video shuts off, you know, I might be annoyed that, you know, I was watching that I'm really into it, but still I'm. So anything that's [00:22:00] latency critical, you will. Always always, always run it over a land or land. You always locally the brakes in your car, no matter how fast that five G connection is and how low latency it is and, reliable it is you will never, ever run latency, critical stuff.

[00:22:14] Overlaps. Always local. so if you're a nuclear plant, you'll probably never connect up to the cloud cause you know, bad news, if something happens, meanwhile, if you're building like no problem, it depends on what you're doing real time in a building automation world is 15 minutes real time, the airbag, a little different.

[00:22:30]so anyway, so that's that when our land are you, in a physically secure data center or are you not. If you are in a physically secure data center, whether it's colo or a traditional on prem data center up in a cloud, that's like Fort Knox, you know, to get into any of those public clouds. It's different than if you're a box distributor out in the wild and someone can put their hands on it.

[00:22:50] Yeah. Yeah. It could be a public right away. It could be, you know, in a, in a closet, in a retail store, it could be your smartphone phone. Of course, that's also part of the continuum. It's not just about IOT by any means [00:23:00] at the edge of continuum includes both client based devices and IOT, gateway servers, whatever.

[00:23:06]so that's physically secure. So in the, as we defined on this taxonomy, so you've got, you know, you mentioned some of them public clouds and you've got the internet exchange and you've got the regional edge, you've got the access edge and it's kind of moving hops close to that boundary condition.

[00:23:18] And then all of a sudden you're on a land, you know, it's the last mile network and then you're on a land. On prem data centers, the first step, but guess what? That still uses pretty much the same tools it's going in the cloud. We are seeing some evolution of those tools, including Coopernetties extending down.

[00:23:32] I've got to manage the trip to clusters, but it is still physically secure, locked out data center and your traditional building, or a micro modular data center sitting outside the building. If you needed some more real

[00:23:40] Matt: Right or, or an

[00:23:43] outpost box or an Azure

[00:23:44] stack

[00:23:44] rack.

[00:23:45] Jason: sure. sure. Yeah. So that's, those are generally, you know, the latter could be even not as physically secure, but generally physically secure, then you get into where, I'm outside of a data center, you know, and the places I mentioned distributed. I don't necessarily, not [00:24:00] only can someone walk up to that box, I got to make sure that you can't load malware on it locally.

[00:24:04] You do not want people to locally log into it. You have to go through some sort of console you want to use rid of tries like PPM and MLA, but you also, you don't necessarily know. What the network is, you don't often own the network when you're just doing, especially as a managed service, depends on who you are.

[00:24:21] You want not like working within a building, so you can't rely on a firewall, you know, per se, like you do in that data center. And so, you know, like what we're doing with Eve with NFV and of course, you know, commercially, you know, with my company, like. It's you have to have distributed firewall capabilities.

[00:24:36] You need to be able to say this app can only talk to that cloud and this outgoing talk to that app and a zero trust security model is part of it. So it's like, you know, only based on a set of

[00:24:48] Matt: We use it. When you say you don't own the network, can you, can you, can

[00:24:53] you, can you, yeah, who's the, who's the you in that, in that

[00:24:55] Jason: yeah. Sorry. So as an example, so we have a customer that we've been working with,  Okay. So if you, if you're [00:25:00] like a service provider and wireline or, or otherwise you on your network, you know, that's, that's a given, but if, if you're,  wireline or if you're a telco service provider and you put CPE equipment on prem, but then, and beyond that, you don't own that network.

[00:25:13] Whatever network they have on prem. We have a customer that we're working with in the healthcare space that does medical imaging equipment, and they sell that equipment as a, release it to hospitals as a service, and they have to update those systems. They're kind of computers embedded inside these really fancy machines, very expensive, and they, they drop AI models on them to look for, you know, we've got X rays and they do not own the hospital network.

[00:25:37] And they, but they have to make sure that their stuff is secure. So they're looking at things like Eve as the, this kind of bare metal of substrate that can help provide that distributed firewall capability. So by proxy, they can only get back to their cloud. They can plant through firewalls and whatnot, but it's still be secure in the greatest scheme of things.

[00:25:52] So it's just a different scenario than if, if you're on prem and you're the network engineering and you own that network, but, We see it all the time. We've seen [00:26:00] other people using, you know, in these just of boxes, you know, Eve as a project, with our controller or otherwise, that was a way to get data off of the network.

[00:26:07] So, cause you know, if you have an OT network, like a process network, that's historically been isolated for safety reasons and uptime, you know, OT actually cares not as much about the security. They do, but what they really care about is uptime and safety. I T cares about data security and governance, you know, and, and whatnot.

[00:26:24] And so that's why when you see these worlds converging, we see a lot of times people are putting these boxes in as like kind of secure network proxy. On top of the process, you can tell an it person real quick. W, you know, the OT people sniff them a mile away. I've seen it. People say, Hey, we could just update from there, cloud your process controllers and whatnot, and just see what really Amelia, and then they're like, get out no, an overlays on top.

[00:26:47] Sure. But you just don't do that. So, but again, it's a continuum, you know, as you go further left, you get into that kind of, well, we called it the smart device edge in the paper. the reason we did that as that, [00:27:00] that, that, you know, so I'm part of an edge. And then you get outside of a fiscally secure data center, but still capable of running apps.

[00:27:05] So the smart device edge is single node with two 56 megs of memory, up to a small cluster with the right features of the cupboard. And that is paradigm, you know, eventually, but outside of a physically secure data center, and this includes. IOT gateways servers, you know, hubs routers, whatever, and smartphones, client devices, PC in the, the client world, iOS, Android, there's various yeah.

[00:27:28] Established ecosystems around, you know, those things marketplaces, you know, windows of course, all kinds of tools and IOT, man. It's the wild West, you know, all kinds of flavors of windows and Linux and different protocols, thousands of OT protocols. And this is why, you know, adjuncts Foundry matters. This is EVE is.

[00:27:44] To do for the, the goal of Eve, Project EVE within LF edge is to do for the IOT component of the smart device edge, the same as what Android did for mobile,

[00:27:56] Matt: And then what is that? What, what did it do? What did Andrew dude from [00:28:00] mobile that's

[00:28:00] Jason: It created scale. It created scale. I mean, Android created an ecosystem and of course you've got iOS and Apple and all kinds of cool stuff, I'm it? You know, I, there are different approaches, you know, I've always kind of a more, well, yeah, iOS is more of a walled garden, but I mean, I have Apple products because I, I liked the experience, but you know, there's different merit to each one, but Android.

[00:28:19] Know, I have people all the time are like, Oh, open source. Why would we give anything to open source? Like open source is the new way to drive standards and standards, you know? And, and people are, I'm like, do you think Google, you know, didn't make money on Android and all the ad revenue that, that ecosystem drove.

[00:28:34] I mean, this is, it's a, it's a, I don't have some sort of open source model going forward. It's going to be difficult because it's ultimately about scaling interconnected ecosystems and that does not work without opening. anyway, so, so. So Android, the Android taped crates at that network effect within the mobile space and creates that open thing.

[00:28:53] And it's a trade off, you know, open it, doesn't always provide the same curated experiences that more of a quote unquote walled [00:29:00] garden, but you know, there's trade offs. but the reason why we put it into. LFS is, you know, to create that vendor neutral governance. You know, there's a difference between open source, just put it on GitHub and open source in a organization like LFS or Linux foundation, where you get structured governance.

[00:29:19] If you cannot govern it with open TSC meetings, the technical steering committee meetings, you know, it's a technical meritocracy. The best way to vote is with your keyboard, right in code. If you don't have that structure governance, it is not really vendor neutral. And, and so we needed it to be vendor neutral.

[00:29:35] So, so is it Nita is a company when you put even into a LFS and we're seeing more and more people kind of pile on. If you install, even on the box, you get, it's an easy button for secure dial tone, you get connected to all of the security benefits, and then you use the absence of choice. We are not in the data path at all.

[00:29:49] And so that gives you sort of like that, you know, Android sorta sorta speak foundation for headless IOT devices

[00:29:56] that don't have a UI in the same sense as an Android device.

[00:29:59] Matt: Got it. [00:30:00] And are there, are there devices that are shipping with Eve as

[00:30:05] standard part of the stack?

[00:30:06] Jason: yeah, there's actually page on the Eve community page called eve in markets. And so we're working with companies like.

[00:30:12] Matt: in the show notes, along with

[00:30:13] the Linux

[00:30:14] foundation, white

[00:30:15] Jason: Yeah. Yeah. We're, we're, we're working with companies like Advantech and planner and Supermicro, you know, broadly speaking Intel, we've got arm, arm boxes that we're adding, Dell, HPE, you know, all of the above and, again, it's like, you know, the Android, the IgG, the one stack you need, and then you'd take your controller choice.

[00:30:32] Of course ZEDEDA would love  to sell you a controller, but we're actually enabling other competitive products because a rising tide floats all boats. We're all better off. If we have more standard components

[00:30:43] and then you win by merit and not lock-in.

[00:30:45] Matt: Yeah. Yeah. And I, and I, and I think that that's the, I agree. I, I encounter people that are like, well, why would I give away my money, my crown jewels. And, well, first of all, you shouldn't get away with your crown jewels. You should always have some differentiator, right? That adds value on top of it. [00:31:00] But w what you, what you want is you want a, an ecosystem of lots of participants that all have a shared interest in building a platform upon which they can differentiate.

[00:31:11] More quickly and deliver value more quickly to customers because if each of us has to go and reinvent, you know, all the layers of the stack for every device, not only do we create in compatibility, which is the pain for customers, in particular let alone developers, but we don't have this sort of shared momentum that we can get by having, by having open standards.

[00:31:30] And I can see it everywhere. You know, my favorite example in this. The younger crowd won't get, this is 35 millimeter film. You know, the camera companies and the film company said, look, we'll compete on the emotion and the materials and the shutter mechanisms and the lenses, but we're not going to compete on the distance between the Sprockets on the phone

[00:31:48] Jason: Right, Yeah. That's that's that's

[00:31:49] Matt: and that enabled anybody to make a film for anybody else's camera.

[00:31:51] And yeah, it gave consumers a lot of

[00:31:53] choice in businesses, a lot of flexibility for competing

[00:31:56] and it grew the overall market.

[00:31:57] Jason: exactly, you know, it's, it's important. We see this [00:32:00] with all the tech companies getting behind like blue Ray and, you know, different formats, of course they're blue. Right. And then, but here's the deal. So this may be, you know, for the younger crowd, you know, remember the VHS tapes versus beta, but here's the example of why ecosystems are important.

[00:32:14] Beta was the superior technology to BHS, but VHS went after the studios and got them to sign on to produce the content. And the content is what wins. And so that's why VHS went out and beta was still used kind of behind the scenes, as a format for like studios and TV studios and stuff like high end equipment.

[00:32:32] But, but you know, it's about ecosystems and you build one join one, but either way it'd be a part of one or you're going to struggle going forward.

[00:32:40] Matt: Yeah. That's really interesting. So you mentioned earlier that you were, you were the part of the founding team of ethics, Foundry. an ethic Foundry and, and Eve, or both, projects within Lennox foundation's LFS. So how do those two projects relate? Like what does edge X Foundry

[00:32:57] compared to Eve and how, how do you use them together?

[00:32:59] How do they even [00:33:00] relate to each

[00:33:00] other?

[00:33:01] Jason: So you're using the Android analogy again. Eva's Eva's more like Android and Jackson's the app. Eva's is a bare metal sort of orchestration foundation with an open API that enables you to deploy things in virtual machines or containers on top of it, your choice of apps, not in the data is that all, you know, you choose what you want.

[00:33:20] You choose what hardware, and the beauty of Eve too is because we can also stick the hardware. I can run on an X 86. I can run an arm. I can run with it. Any coal processing. So last I checked, it's not just about GPU and FPGA is, it seems every day, there's a new code processing, Silicon company being announced, CPU, unit, whatever.

[00:33:37] Yeah. Any of that. So you need to be, you have complete choice, Raul edge X, then rides in the application layer above, you know, you know, so edge X is about it. A lot of people think, you know, protocol support, So, so like pick my bus legacy protocol, use widely and energy buildings, you know, whatnot.

[00:33:57] Backnet same thing for building automation. [00:34:00] mod bus has two versions. One runs serial over serial ports are 45 and one runs over ethernet there's a certain degree of hardware support required.

[00:34:12] Yeah. When some people don't realize is that there's, there's lower level supporting cereal does not mean you support my bust.

[00:34:20] It means you support the transport, but there's an application layer, protocols. So it's almost like the, the telephone wires then the people talking and then the dialect, but specific dialect you use does that last little bit. And that's what edge X does. When, when that power meter, you know, what does it mean?

[00:34:35] What's the semantics for voltage and, and, and, you know, current and whatever, today of 400 and something IOT platforms, every single one of them writes that protocol different, you know, that support for that one power meter. So I have 450 drivers for one device

[00:34:52] Matt: so adjunct sort of normalizes all of

[00:34:55] Jason: Yeah. It's, it's a universe. It's a secret. Right. Yeah. [00:35:00] Yeah.

[00:35:00] Matt: these different

[00:35:02] Jason: Yeah. Think of it as a double translation engine. So you write protocol drivers that plug in on the bottom, open API in the middle, and then you write cloud connectors or backend application services on the top. And an ethics is about those APIs in the middle.

[00:35:15] And so as long as you plug in, you know, and you can crowd source all that, since it's a microservices architecture, like you can write a protocol driver for your sensor and plug in that the, the sensor makers don't do that today. Cause I'd have to write 450

[00:35:27] of them for on

[00:35:28] Matt: So, so let me, let me, let me see if I got this right. So let's say I make a sensor that allows you to read temperature.

[00:35:33] I imagine ethics. Foundry has a general. Semantic concept of temperature, give me the temperature. so I would take my device in the microcode and the software that's on that device. I would then I'd write a driver that plugs into edge X Foundry that says when you get the request

[00:35:53] Jason: Yeah, and, and, and the SDK has made it to where the way I talk temperature, my dialect is [00:36:00] now translated to the, the kind of secret decoder ring and edge X. And then that transfers it to whatever language that your North end wants to speak towards the cloud or to wherever you're going. And, and so it, have you ever watched big Lebowski?

[00:36:12] I would joke it's, it's the dude, it's the rug that ties the room together, you know, so to speak, that run really tender, but also, You know, it's just, and we were very careful when we started that we did not want to pick a data model in the middle. So it's very flat in the middle. Just Switzerland is data passes through, and then you choose whatever you want on the other end, because I want you to pick a data, model you wrong to somebody, and then they don't want to use it.

[00:36:33] So we were very particular. And then the reason why X is in the name, it's because it's trademark. And so you can have X compliance, sensors, and NetJets compliant apps and cloud connectors. And so this is about creating an ecosystem, but all of that rides on top of there's a, there's a program spinning up within the adjuncts,

[00:36:54] Matt: That's another reason for it to be

[00:36:57] Jason: right?

[00:37:02] [00:37:00] Oh, yeah. Well, and you know, and open source in general has just become the, the it's the modern way to drive standards. It used to be that I get a bunch of people together, Geneva, or otherwise we talk a lot. We write a document and the people do stuff to that document. Well, it turns out I can run, run code.

[00:37:17] I can't run, you know, PowerPoint or, you know, whatever, like a document. And so. Collaboration and open source. It shared technology investment. As you've mentioned, I picked up this word, I can't remember from who, but, like it minimizes undifferentiated, heavy lifting. Stop doing that start focusing on value and, and it, it helps you create a snowball effect for standardization standardization with specs is true STO standard, bodies, you know, writing specs, that's the truest form of standardization.

[00:37:47] But what open source is treating is the fact of standards. If you get enough people using something, it's just a fact of standard. And, and it's just a new way of, of driving that sort of network effect for standardization

[00:38:03] [00:38:00] Matt: So, you know, one of the things that, that people talk about a lot is, The role of artificial intelligence. And, you know, obviously with all these sensors, creating all this data, the, one of the best ways to process it, not the only one, the best by the processes is with AI. So how do you see. You know, this, this other emerging massive trend, which is, you know, I mean, AI has been around forever, but not in the way that we have it now where like machines can actually teach themselves some pretty amazing, techniques, like identifying faces.

[00:38:34] So how, how do you, how do you see edge and AI and the

[00:38:42] Jason: Yeah. Yeah. So AI, obviously a big trend and, you know, as a subset of machine learning and you know, but I'll also say that I've kind of, there's a lot of people out there that, are doing basic rules. If this, then that rules engine said, look, look at my AI. I'm like, no, no, no, there's a, there's a guy. I was.

[00:39:01] [00:39:00] Yeah, yeah. Yep. Yeah, there's a guy on the line. I just need that attributed it. I can't remember the name, but, did a post that went a little viral. It was like, if it's, if it's machine learning, it's probably written in Python and if it's AI, it's probably written in PowerPoint.

[00:39:20] So, yeah, but, but, you know, no AI has definitely come into the zone. I mean, a lot of this coal processing, a lot of the, the compute power is helping, you know, these building, these neural networks, there's standardization happening, things like Onyx, you know, helping you kind of bridge standards, of course, TensorFlow, you know, really picking up.

[00:39:35] So a lot of AI happens initially in the cloud. I mean, that's where the, the deepest of the learning is going to happen. You know, that's a scalable compute, but then as people start to realize. You know, this, the bandwidth problems, latency issues, privacy issues in terms of data, sovereignty issues. this is why you're seeing those tools extending to the edge.

[00:39:53] And, you know, like we said before, it's a continuum you need to be to prepare for AI world. You need to [00:40:00] be investing in a variety of abstractions. so starting at the foundation, I mean, this is why Eve matters. Why projects like Ajax, cause a lot of good stuff out there. Of course it's not just about those.

[00:40:09]any really LF edge in general. You want to the last thing that you virtualized great as it should people second biggest assets, your data, and the last thing, the virtualized is data. And if you separate data out from applications and then of course, applications out from infrastructure. Then you, you can, and you make the right investments in portability across that continuum.

[00:40:31] I don't need to know today, you know, where I'm going to be running my AI models. I can kind of move them as we go. You're going to start in the cloud and certainly your training. And then maybe I deploy an inferencing model to the edge because you know, like I said, with computer vision, you know, I want to send events.

[00:40:46] There's a car park there for four hours. That's an event versus. Here's a video of a car that just ate up your server, your network bandwidth and. You know, so that's important. So ed Edge AI, you know, the killer app, it's computer vision, high [00:41:00] bandwidth, same as, you know, using AI to analyze for that little tick and the machine with, you know, vibration data.

[00:41:06]if I'm running a thousand Hertz or 8,000 Hertz, a kilo Hertz of vibration data to look for the little needle in the haystack that says this, this motor is about to fail and, you know, a month or two or a weeks or weeks or whatever, I don't want to send that high bandwidth data over the web. You know, I want to analyze it and say, Hey, got a problem, send a partner at tech.

[00:41:26]so we're seeing that happening. And so AI, so number one, it's it's. It's a con it's part of the continuum, you know, edge AI. we also see, of course it's about investing in cloud native principles everywhere you can. So this notion of containerization and platform independence, and, and you do not want to hard couple, any given edge data source to any given backend.

[00:41:48]because then you're kind of locked into these silos you need to kind of, so there's all these layers of abstraction. but you know, the key considerations for Edge AI standardization. how do I deploy all those models and make sure that they're not deviating? [00:42:00] how do I prevent false positives?

[00:42:03] They're just, you know, detection,  go Google, Chihuahua, AI muffin, you know, online and you'll see like pictures of like, you know, AI messing up. It looks like a blueberry. It looks like they said, this is a Chihuahua, but actually it's a blueberry muffin, you know? Can I tell the difference between a car, a bicycle and a lamp?

[00:42:20] Sure. But it's a little harder when you get into some of those weird things, ultimately though is with AI. it's about domain knowledge. Data scientists don't have the domain knowledge. They just know how to program AI, those tools, the algorithms for things like facial detection, license, plate detection, you know, how old is that person demographics?

[00:42:41] Is it a car or a lamp or a bicycle that's going to become commodity. They're already kind of becoming commodity. Cause people are exchanging data, you know, training, you know, things. Longterm AI will be about creating models for very specific contexts. You know, that part geometry on that line for quality control, you know, [00:43:00] this interface for that, you know, thing, this particular medical field.

[00:43:04]and so that, that's where I, I would caution people that are getting into AI. It's a great, great space. I mean, it's super, super important, but at the same time, Don't do undifferentiated heavy lifting, you know, go after the, and that's why you're seeing open source also really bringing up, you know, the AI thing and, you know, people just getting to good data is, is a Trek in the U S especially when you have to watch for privacy and all that, there's other regions of the world that don't really care about privacy so much.

[00:43:30] And so they're ahead on it. Yeah. Hi, because they've got massive sets of data that they're training their algorithms with. And, and so, you know, it's all. Part of it ethical AI, big consideration. But knowing that that model is deployed and you don't have weirdness with the camera angle changing and the lighting conditions bad, and all of a sudden you got a Chihuahua instead of a blueberry muffin.

[00:43:50] Matt: Right. Yeah. You know, one of the things that, you know, people often ask me, what do I see as the drivers of edge computing? And there are a lot of drivers and it's a, it's a, you know, [00:44:00] Pretty complex emergence. but one of the big drivers is we're moving from a world of humans, talking machines to machine talking machines, talking to machines.

[00:44:08] And one of the things that people that aren't, down in the weeds like you and I tend to be, don't tend to realize is just how many, how much the scale changes. It's not just like things get a little faster. The data gets a little bigger. Right. if you think about. A billion sensors, generating petabytes of data.

[00:44:27] A petabyte is 1 million terabytes,

[00:44:30] Jason: Right, right, right,

[00:44:31] Matt: right? Not enough fiber that it's faster to put that on a disc drive on a train than it is to ship it over, over the internet, given you know, a very large pipe. and then you think about like time, right? I mean, humans operate in ones of seconds or, you know, maybe free fractions of a second.

[00:44:47] If you're looking at the. Yeah, exactly. I have to say it. Yeah. Yeah. But, but things, most of the things that we're consciously aware of happening in a second ish, and you know, machines operate in microseconds in [00:45:00] nanoseconds and a microsecond is one, 1000000th of a second. And so if you're trying to process, you know, many.

[00:45:08] Terabytes or petabytes of data. And, in, you know, a few millions of a second, you're going to have to do it well, first of all, much closer to the device. And you're going to have to do that with tools that are only now coming into calm, common existence, which is, you know, servers that can handle that kind of data in real time.

[00:45:27] And you know, one of the examples you brought up earlier was the autonomous car and. I don't know, maybe autonomous cars are coming out. It was sort of going out favor because we all thought they were going to be here in 2020. And

[00:45:38] I think the dust is settling and I realize there's nowhere. We're a decade away, certainly for urban navigation.

[00:45:44] Right. Yeah, exactly. Exactly. You can see the world change, but I think what's interesting is, you know, today's autonomous cars for the most part, have all of the processing on the car. They've got all the sensors

[00:45:55] on the car and the car truly is autonomous. It's designed to be disconnected, frankly, because there's no [00:46:00] network support it.

[00:46:01] and that's also the best way that they can, you know,

[00:46:04] at that scale collected and process the data. Cause you just, you basically put a data center on the car. now you don't have to think very hard about that to realize. That that is not the most cost effective way to deliver autonomous driving to the masses.

[00:46:16] Right. You can't put a quarter million dollars for the servers and a car and charge $20,000 for it. so obviously all of these things will be sized reduced and cost reduced, but, okay. So there's a set of functions that can improve the experience that can happen. You know, in the cloud, so to speak probably an edge cloud, so that it's low latency, but they don't have to have it on the car.

[00:46:37] I mean, obviously you want the braking system that stops you from crashing to happen on the car, but let's look at an example where I'm approaching an intersection. My lidars. Can't see around the corner. There may be other cars coming or pedestrians or some other object coming. If my car is talking to a nearby server, that's looking at traffic flows and people from the directions, it can tell my car that you've probably [00:47:00] should start gradually breaking.

[00:47:01] So even though the car would still stop me safely, when it could detect the thing around the corner, hopefully, it might cause me to. You know, it might cause a very uncomfortable ride. whereas if you've got some intelligence. So, so it's really interesting how, you know, with, with cost and, features and capabilities that we're going to be making these decisions about where we place these workloads, because some of them can run on the device and should run on the device and some of them can run other places.

[00:47:33] I'm just interested if you're

[00:47:35] Jason: you're going to see a mix.

[00:47:36] Oh yeah. You're going to see a mix if it's okay. It's like, if it's latency critical, it's going to run on a land. So the land in the case of a car is the car. And then if it's latency sensitive, like approaching the intersection, you know, but I can augment, you know, the decisions that each car makes.

[00:47:49] Plus the person that's looking at their phone about to walk in front of that car at the intersection. Maybe everything there's like a [00:48:00] new trend or whatever,

[00:48:01] something like that,

[00:48:01] Matt: Yeah.

[00:48:02] there's vehicle to vehicle networks and vehicle to infrastructure networks, and probably both are going to exist and yeah,

[00:48:07] Jason: what you also need open standards for, by the way that anyway. but yeah, so you're going to see that. So infotainment, anything around augmented reality for heads up displays things that augment your experience. You always want to run. But more often than not, you want to run, further up the stack at the access edge, regionalize, whatnot, you know, some colo location, like what you guys do, but like, you just, you're going to see a mix and it really, you know, me and my privacy.

[00:48:31] So a lot of people to think, Oh, it's all computes happening on servers. Well, guess what, increasingly it's happening in constrained devices, you know, that lasts the edge as part of that continuum, tiny machine learning, tiny FML is a big trend just over the past six months. There's all kinds of talk about it.

[00:48:44] Easy example. Hey, Alexa. You know, in theory, Alexa is not talking to anything. I mean, he basically likes it doesn't and was like all worried about it. It doesn't have any internet connection until you say, Hey Alexa, and there's a tiny algorithm inside of a chip that says, Oh, wake word. Okay. Now make the connection.

[00:48:59] And [00:49:00] so that's a good example of tiny ML. So you're going to see more of those very fixed function, things happening everywhere in general, too. AI is today. It's a little bit of a misnomer because true AI to me means you have emotion. You've got morals and ethics and all that. That's why the human, you know, that's a big distinguishing factor, a big distinguishing factor for people.

[00:49:20]but AI today is more about. Can I program a very complex set of roles into a model, like, you know, speed, you know, distance road conditions, lighting conditions, everything, and for a car breaking in an intersection, things that are constrained by a structured set of rules, but then there's the old, you know, the thing about like, Hey, do you run into this set of people that look like a bunch of thugs?

[00:49:43] Or do you run yourself into a tree or do you get the, you know, one non or like whatever, you know, none of that. Yeah, I think we should have perhaps decide, but we see those types of decisions at AI. It's just not, you know, that's not the same thing. I don't know something you said earlier just reminded me.

[00:49:58] so a lot of people are kind of in [00:50:00] this innovator's dilemma, you know, right now   just in terms of how do you. How do you pivot to change? It's like I do a lot of conference stops as you D do you want me to be not so much, you know, it's a lot more virtual these days, but still.

[00:50:12] And when I talk about technology to people that are just sort of in a different industry, like maybe it's the, I'm talking to a bunch of people that do the building maintenance or something, and they just have their, their world and the ITP, my very world often I get the munch of people staring at me with the arms folded.

[00:50:26] Like, I don't like you because you're talking about things that make me change. And then you throw the Uber logo up and say, look what they did the taxis in six months. Yeah, the whole industry. so I was on it. Just remind me about this. I was on the phone with him. Oh, it was the connected machines, you know, more and more machines being connected.

[00:50:41] So I was on the phone with a very large payments processor. I'm not in the name. you know, they do like the payments, sores and whatnot, credit cards and stuff. And I knew that they'd gotten really impacted by, you know, people like that, you know, to completely change the game and if mobile payments and I'm like, I'm like I thought about this ahead of time.

[00:50:57] So I get on the phone and they're trying to figure out their IOT strategy was [00:51:00] clear that they're just like, I don't know. I don't want another square to happen to us. I'm like, Oh yeah, I can imagine that that really kind of impacted you and you, you want to go figure it out and there's always kind of what to do and, you know, kind of stuck in the innovator's dilemma.

[00:51:11] Like if we just do what we're doing, we've always done better. And I planned this for the end of the call and I said, Hey guys, I gotta go. But have you thought about when machines start making payments, you know, it blew their minds. And I'm like the fact that I didn't say this, but the fact that that blows your mind is exactly why I know there's queer is going to happen to you.

[00:51:27] Yeah. You've got to get outside of your bubble and start to think about how to pivot. To an adjacency and kind of a machine doing analytics locally, I'm about to break. And then I'm going to order apartment a tech, but guess what? That machine is talking to another machine across the network, and then you're coordinating.

[00:51:45] So don't just send out the tech to that machine. This machine is going to be break two in the same region. So drive out once and replace that both parts. You know what this machine is so broken, you've replaced the parts so much that what replaced that machine, it's going to cost you less over three years.

[00:51:58] These are those [00:52:00] interconnected, you know, decisions that you start seeing. We're not at that point yet as today. It's just like what's going on with my machine. But over time, this network effect across ecosystem, that's just for your business. And then imagine all these other businesses, interconnecting and retailer crossing them in the home and all that stuff.

[00:52:15] This is the true power. That's why the last thing, you know, I had done before coming over to,   video worth, continuing to work on it, you know, with, with the of folks is, you know, announced this project called ovarian, which is around, this notion of data confidence fabrics. How do I send data across networks with measurable confidence, long story, you know, but it's basically how do I start to enable what I, what I've always kind of called them.

[00:52:36] Holy grill, digital, which is basically maintain privacy and IP protection, but sell stuff to strangers. Data resources and services to people I've never met ultimate scale, and you can not. I take people out to dinner fast enough to build the trust. You must have technology to help you. And it's not just about it's all of these layers, right?

[00:52:52] Root of trust, open API APIs, immutable, storage, ledger, confidential computing system level trust. And that's what other learned about. So there's some, there's a really cool [00:53:00] video that we produced online and nobody rammed up word, but.

[00:53:06] Matt: Yeah. Well, I mean, we're, we're almost at the end of the interview and I kind of wanted to ask you two more questions that, that, we'll we'll have you draw on your, your long. The length of time you've been in, in edge and IOT. and so I guess the first question is if you think about, you know, the, the, the arc from

[00:53:36] Jason: I think the biggest change, aside from just people starting to realize where the right use cases are versus, Hey, isn't that neat? But just because you should, it doesn't mean you should or does it cause you can doesn't mean you should. I think the biggest change is people haven't gotten out of their systems.

[00:53:49] That it's a good idea to try to own everything. What happened, why we got to four 450 platforms? Is it everybody rushed as you build all those platforms, but [00:54:00] really when it's Ultima ultimate, it was the running joke. I'd probably pick that 500. I don't know when we started execs, we actually fought internally at Dell.

[00:54:07] We're like, Hey, do we make our own platform? Then I'm like, no, I don't want to be the 201st platform we joined. We decided the team that they seem to be open. You know, you know, the long history of open stuff. I mean, this is why that's one of the main reasons. It's a great company, but still it's like, we, we don't need all these platforms. We need necessarily unique hardware and software, and we need people offering services and people with domain knowledge, the biggest thing that I've seen in why we started as X, Y w I help, you know, with U shape LFF edge, and a bunch of great people is that we've gotten to get away from everyone trying to do everything.

[00:54:36] Cause they're all like I'm going to be in the data path. I'm gonna own everything. It's gonna be awesome. And then Holy crap, that's hard. We need to band together and maybe I should do one thing really well. And so the way I'm building an ecosystem with the team at ZEDEDA is I want an entourage of problem solvers that plug into a more of an open heart or model, you know, really good experts that augmented our own security benefits.

[00:54:55] Really good ad company is really good people that do clouds. I mean, we're working very [00:55:00] closely with Microsoft and others. For example, integrating with those tools,  where the hardware providers, it takes a village. Anytime anyone does everything, rarely do you do one thing well. Thats the biggest change ive seen over the last five years is people realizing maybe i should have a go-to dance move

[00:55:16] Matt: Yeah. Well, and, and I certainly am feeling that too, that sense of like, okay, a lot of these things that felt loosey goosey a couple years ago are starting to get traction and meaningful collaboration and meaningful interfaces and meaningful products. And so I guess looking forward, if you could wave a magic wand, like what's the, what's the one domino you would topple to accelerate,

[00:55:45] Jason: Oh, man, that's a tough question. I think the next, I mean, one of them is just continuing on this interoperability thing, but I think the next big domino to topple of why we kind of tried to get ahead of the curve a little bit, you know, as a thought leader at Dell and while we're working on it, there's [00:56:00] other great efforts.

[00:56:00] Like the trust over IP foundation, there's all this, the next domino that you have to solve as trust. You've got to figure out a way to scale trust over heterogeneous networks or none of this stuff will ever work.

[00:56:15] Oh yeah. I went now. I mean, we've got, we've got a path and if some of it's just going to be an in building.

[00:56:22] Matt: out of, out

[00:56:22] of?

[00:56:22] Jason: okay. It's going to come from a variety of different consortia efforts, open source collaboration. I mean, we put out very Aminta as a Linux foundation, you know, it's a seed project. It's so kind of bootstrapping and all that, but. Al varium, you know, Alverno.

[00:56:35] Matt: Yeah. we'll put that in the show notes too.

[00:56:37] Jason: it's Latin for beehive.

[00:56:38] Like interconnected hunting comes with the video is really cool. It shows how like a day in the life of how you go across all these things. It's exactly. It has the intersection sharing cars, sort of blah, blah, blah, you know, talking to each other intersection and it shows retail is coming into the home.

[00:56:50] There's a reason there's a service provider in the middle of it because service providers massive opportunity to become a service provider for service providers. Aye who owns the trust in the home with today? I mean, [00:57:00] I mean, I know my ups driver by first name that goes on Amazon. Amazon's great. But how do you change the game?

[00:57:05] You make it to where nobody owns a trust. Imagine a home gateway that you could drop in retail apps and health care and insurance. Right. Yeah. And there's no way that those things can cross pollinate the apps only when you set the terms and you know, so they'll vary. And video goes through all this stuff only based on privacy terms.

[00:57:23] You said in generally speaking as consumers, if you get value and you trust it, your privacy you'll give up some privacy. And, you know, if I told you 10 years ago, you would leave location based services on your phone. You'd think I'm crazy, but people, most people do today because they get value, whereas in the aerospace to join or whatever.

[00:57:41]you know, this is going to evolve, but th the, the biggest domino to fall for that is number one, it's about interoperability. Number two, it's about building trust at scale, by using technology to help you. you know, and that's what, Alvarium's about, these other things that's much further out. But the thing I always tell customers in my role is so I build ecosystem.

[00:57:57] It just helped with our , leadership. [00:58:00] You will never, well, it was scaled to the grail. I needed a rhyme. I actually, I got a slide. I joke. It's like shake and bake and Talladega nights. When I shake and bake, you know, scale and grail, you'll never scale to the grail. If you don't take an open path, even if right now it's about simple solve, some simple problems, create some value.

[00:58:16] In increasingly, you know, be a part of an ecosystem, see this network effect, but you'll never get to the ultimate goal of, of just entirely new business transformation, across many markets. And this just sort of mashup of everything, if you dont have an open base. That's why we're all working on this stuff, and Then let's And go make some money around the wheel.

[00:58:42] Matt: for joining us today. It's been a great show and how can people find you online?

[00:58:54] I love it. And, and the, the, the listeners can't see this, we're on zoom recording this, but,  Jason's sitting in a room [00:59:00] surrounded by instruments. I think I count

[00:59:01] at least three guitars and a drum set for, we get to five avatars.

[00:59:05] Jason: Yeah. So I play music. You can find me online too. As a, as my band, I'm not going to, you know, really promote it, but easy to find. But, people always ask about my

[00:59:12] Matt: I love it.

[00:59:13] Jason: office is my home city and people are like, where'd you play music? I'm like, no, I'm in Austin.

[00:59:18] When you move here, they just give you all this.

[00:59:20] Matt: Right. This was the last owner left

[00:59:22] it. He was moving out of Austin. He can't take

[00:59:23] Jason: Yeah, it's a music town. Just take this stuff. You want a drum set too? Here's a ukulele.

[00:59:27] Matt: deaf shepherd on, on Twitter. obviously you're involved in Linux foundation. People

[00:59:31] can

[00:59:31] Jason: Yeah. Easy to find,

[00:59:33] you know, a board member there as well. And so I'm easy to find online, generally speaking and whatnot.

[00:59:37] And you know, I'm always up for a conversation, you know, it's just, like I said earlier, this is about collaboration. It's about surrounding yourself with good people, you know, at the network of fact and you know, everything else follows.

[00:59:49] All right. Have a good one.