Today’s episode features an interview between Matt Trifiro and Rodney Richter, Enterprise Architect at Hewlett Packard Enterprise and member of the LF Edge Technical Advisory Council In this interview, Rodney takes us through the technological and logistical challenges of implementing edge solutions at the enterprise level, the thinking that led HPE to develop an entire product line of edge servers, the types of use cases that are getting traction today, and more.
Today’s episode features an interview between Matt Trifiro and Rodney Richter, Enterprise Architect at Hewlett Packard Enterprise and member of the LF Edge Technical Advisory Council
Rodney has been with HPE since 2015, and previously held numerous positions over 19 years with AT&T, including Principal Network Planning Engineer responsible for AT&T Wireless and Wireline access network virtualization.
In this interview, Rodney takes us through the technological and logistical challenges of implementing edge solutions at the enterprise level, the thinking that led HPE to develop an entire product line of edge servers, the types of use cases that are getting traction today, and more.
Key Quotes
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.
Links
[00:00:00] Matt: [00:00:00] Hi, This is Matt Trifiro, CMO of edge infrastructure company, vapor IO, and co-chair of the Linux foundation, state of the edge project.
[00:00:06] Today, I'm here with Rodney Richter enterprise architect at Hewlett Packard enterprise and a member of the LFEdge technical advisory council. We're going to talk about Rodney's career in technology, his work on mobile networks and his current role where he works on HPE's joint program with AT&T and the future adoption edge solutions.
[00:00:24] Hey Rodney, how are you doing today? I'm doing terrific. One of the things that I like to find out from people before we start talking about edge and all this detail stuff, just how, how did you even get involved in technology?
[00:00:36] Rodney: [00:00:36] my technology goes way back. Uh, probably when I got out of high school and I joined the air force. So I, I always had a knack for technology and electronics, and I joined the air force out of high school and got involved in electronics and supporting all the systems on, uh, like F. F 16 aircraft. Um, and I was part of a F 16 fighting squadron, um, in Southern Georgia.
[00:01:00] [00:00:59] And everything from RF technologies to lasers, to, uh, terrain following it was train following radar, and then forward-looking infrared all those kinds of systems. So that kind of kicked off my career. And naturally I just wanted to continue that in electronics, got my degree in electronics and then eventually, um, ended up with, uh, one of the, uh, the major telcos at and T.
[00:01:21] Matt: [00:01:21] Yeah. And, uh so how long were you at at&T.
[00:01:25] Rodney: [00:01:25] Uh, about a 20 year career there, I started in 96 and, uh, went through until 2015.
[00:01:30] Matt: [00:01:30] And what, what were your primary roles
[00:01:33]Rodney: [00:01:33] Uh, primary roles. Um, as I, um, look back, it was starting off in what was building their long distance network with one of the local, um, operating companies at Meritech back in the day. And, um, so that was supporting all the functions of, the old telco, uh, calling, um, Information systems and, and then eventually got into computer based support [00:02:00] servers, storage, networking, um, supporting customers, internal customers within the phone company.
[00:02:06] Um, and then I also got involved with, um, Network planning and engineering at, at and T for about eight years supporting video platforms within a, at and T and then, um, some software defined networks and looking at, uh, what was at, and T's domain to effort to move from legacy systems, to, their new technology software defined, um, looking at, how do you evolve your networks for the next technologies, especially moving into where we're at today, 2020 and beyond.
[00:02:36]Matt: [00:02:36] It's, it's interesting. You and I, uh, are probably close to the same age. And, I grew up with rotary telephones and, uh, physical switches and, we were joking before the. For the podcast about dial tone. Those are actually a little machines that sat, sat on the local loop and generated a tone that you knew you could actually dial, uh, and which is such a different [00:03:00] world.
[00:03:00] So you were right there in the transition to IP based networks. What was that like?
[00:03:07] Rodney: [00:03:07] Yeah, that was a, that was interesting was at the time, today we're all, there's a lot of openness, about, being able to do a lot of collaboration do look at, what things can be created and developed in an open, uh, community. Um, back then, it was, there was a lot of taboo about openness.
[00:03:26] And when you look at the evolution from those old networks to, to the IP based, um, we looked at, we wanted, uh, resiliency. We wanted to ensure that. Everything was going to work as expected. Didn't, didn't want to bring in a lot of the openness at the time, but, um, moving to IP networks was a challenge because you had to look at how do you evolve from an analog world to a digital world.
[00:03:49] And, and nowadays, a lot of it is moving towards, obviously we're in an IP based world, but now we're in a technology world where, how can we leverage community [00:04:00] to be able to expand what's capable. As well as be able to bring that in and, and look at, what technologies and offerings can bring value to the business and to our customers.
[00:04:13]And by having that open community really is, is a big
[00:04:17] Matt: [00:04:17] Yeah. I, I, I hadn't really thought about that. Source on the scale. The time scale that we're talking about is relatively new. Um, I mean there was, Richard Stallman source, which is everything should be free, which is very different from, Linux foundation, open source and modern, modern open source.
[00:04:34] And I can see the, the, I, I worked for software companies where the idea of giving away your software was just completely crazy. Um, and I imagine like you tried the, the, the. The it, the proprietary, everything, uh, was very hard. So what, what was that transition like? Like what, what drove that transition and how long did it take and where are we now?
[00:04:57] Because at, and T is a, is a, is a [00:05:00] big contributor to, to open source, uh, part of the LF edge, a big part of the LF edge and a bunch of the other open source initiatives and obviously Hewlett Packard enterprise this as well. Can you just give us a little history lesson on. What it was like to transition to open source and why it was seen as important, even though it was a little bit of a, is the body going to reject the organ transplant?
[00:05:23]Rodney: [00:05:23] Um, let's see for, for the transition. Um, it's probably been kind of a, a slow adoption if you will, because if I look at my 20 year career, At and T and, and looking back, as I mentioned, things were a little bit more taboo. If I had a tool or something that I wanted to use, um, to be able to monitor something, and it was an open source tool, um, could I bring that into the network?
[00:05:45] Was it a security risk was, that mentality, um, way back when, and there's still always concerns over security, but if you look at, and I'll, I'll take that 20 year span, it progressed year over year, where. How do [00:06:00] you adopt new technologies and how do you drive cost savings into, your business and with proprietary technologies?
[00:06:10] There's an inherent cost in doing that. Obviously you have the, um, the assurance to some degree that you're going to have. Things that are going to be, um, secure. You got one vendor you're going to be dealing with, if it's a proprietary solution. And that that software is going to work on that hardware, et cetera.
[00:06:28]Um, But in doing so you're somewhat locked in. So to, to move to an open source model gives you the capability to evolve that, to, um, driving down costs. Um, and so getting back to the history or, or, that that 20 year span is I've seen it go in probably about five year increments, I would say, where.
[00:06:50] Originally it wasn't, very well adopted too. Okay. Let's bring a little bit in over the next few years, let's test it. Let's make sure it's secure, get a comfort level around it. [00:07:00] And then probably within the last 10 years is where we really saw the explosion of what's happening more and more involvement in the open community.
[00:07:09] Um, as you're well aware at and T is involved in things like a crane, um, and other areas and really a, um, A sounding board for other industries to be able to, uh, leverage and use open source to help drive that adoption. Um, I've been seeing that as, as I'm part of the LF edge TAC and, um, w we also see what's happening in a crane now working with at and T.
[00:07:34]Um, that, that's where I see it's been, has been over the last 20 years.
[00:07:39]Matt: [00:07:39] So I haven't talked about the economics of shared infrastructure and I'm typically thinking about, buildings and fiber optic lines and networking equipment and all of that. And, and, but there's also, I think, a direct analogy with, with open source.
[00:07:53] Um, and I think other industries too, you think about. If I'm a trucking trucking company or a, a company [00:08:00] like FedEx delivers packages like. I'm unlikely to see any benefit from building my own roads or my own airports. Um, and all my competitors probably feel the same way. And in those cases we had the government come in and do it.
[00:08:15] Um, but a lot of times the government isn't as fast as we need. And so it seems like a lot of these entities, uh, many of them. Source, but also some of them private, like my, my company's business is private, but it is very much neutral, neutral hosts, neutral carrier. Uh, but the idea that you can leverage work from other people collectively, but still create enough differentiation on top.
[00:08:36] Uh, it's a very powerful concept and, and, um, I think it's, I think we're going to look back. In a hundred years and it's going to, it's going to be a major contribution to the accelerated technology. Is, are these, these common platforms that we've all agreed to to share? And that the patent system was a little bit like that, right?
[00:08:54]I'll tell you how I did something, but I get an 18 year exclusive and open source, just recognizing [00:09:00] 1818 years is, it's, it's, uh, uh, a lot can happen in 18 years. And so we actually need other systems. That's, that's a very important, but we'll come back. We'll come back to shared infrastructure and business models, because I know that that's something that HPE has spent a lot of
[00:09:13] Rodney: [00:09:13] Sure. And one point I'll make, just, I think, on the open source side, it's more of a compliment to what's available. And I don't think it's an end all to, every, um, uh, an entire solution. And you see a lot of. Of, um, proprietary solutions that have plugins or APIs to be able to plug into various different, uh, applications, et cetera.
[00:09:32] So I think you're going to see, uh, a hybrid approach to two solutions.
[00:09:37] Matt: [00:09:37] Yeah, that, that seems to make a lot, a lot of sense to me.
[00:09:39]So tell me about the transition from AT&T to Hewlett Packard enterprise.
[00:09:44] Rodney: [00:09:44] Sure. Um, it was probably in 2014, 2015 when I was. Just looking to, uh, do something a little different. As I mentioned, I spent about 20 years, uh, working for at and T in various capacities and, um, [00:10:00] decided to join Hewlett Packard enterprise. I had moved from the Midwest down to the Southeast, so that was a bit of a, a change for me.
[00:10:07] And I did spend about three years in the air force down in Southern Georgia. So moving to, uh, to Georgia, um, back in 2015 was, it was a bit of a. A change just from having to move across the country with a family at the time. Um, that was, the kids were somewhere in high school, summer, middle school, and moving from more of a, in my job role, which at, at, and T was more of, uh, architecture network planning into joining the sales team at Hewlett-Packard enterprise on the at and T account.
[00:10:38] And so I had never been in sales before. But I was I've, I've never had any issues doing any, any type of position. I, I felt like I could always, um, learn anything and, it's, there's certain skills that just as a person, being able to interact with different people, um, that's, that's something that can't necessarily be taught.
[00:10:59] And I always had a [00:11:00] knack for working with people, understanding technology, understanding business, and being able to, uh, to do that. So moving into Hewlett Packard enterprise. Coming into the account team and working with. Um, what was my colleagues? I've worked with a bunch of people at, AT&T.
[00:11:16] Um, when I was at, at and T and now they were my customers. So that was a definitely a transition. Um, it, it was great to be able to have those relationships. Um, it just, it changes the dynamic. When, you go from being a colleague to now you're. You're a supplier vendor, a solution provider. Um, but definitely, um, good relationships still exist.
[00:11:37] And, um, it, it was really a learning challenge for me for the first two years, getting into sales for the first time. So that was, uh, that was good, but now I'm a seasoned sales person. Um, and I'm, I'm a technologist within the sales team. So I focus a lot on the technology.
[00:11:54] Matt: [00:11:54] So like S solution architecture and technicals cells and things like
[00:12:00] [00:11:59] Rodney: [00:11:59] Exactly. Looking at, what's the infrastructure, what's the applications. What, how do you, ensure that things are gonna work from, uh, an end-to-end solution, um, taking more of a holistic approach to, if somebody says, Hey, I need, uh, X amount of servers to serve, uh, an application. We want to look at it from.
[00:12:19] Like I said more of a holistic approach of, okay. What does that end to end solution look like? Cause if you place a piece of compute into the environment or, or a piece of storage, Unless, what's happening. And to end, you may not know exactly what's required. Are there going to be security protocols that need to be put in place that you might not think about?
[00:12:37] Unless that application is, is connecting from point a to point B, how do you have to segment your, your network? Uh, do you need to look at other connections? What are the customer's security measures they want to put in place that might impact that the infrastructure you're putting in place, all kinds of different variables come into play on that.
[00:12:56] And that's really where I come into into the role to be able to [00:13:00] talk with the customer, understand what they're looking to do and help them develop that solution as well.
[00:13:06] Matt: [00:13:06] Yeah. And I realize obviously some of this, uh, you won't be able to talk about in detail because of, uh, the confidential nature of the relationship. But I'm wondering if you can walk us through, um, an example or two of some of the problems that you were presented with and how, how you. Uh, identified the end-to-end problem and delivered solutions to that.
[00:13:29] Are there any that you can talk about, or even if you have to genericize them a little bit?
[00:13:34] Rodney: [00:13:34] Yeah, actually. Um, let's see if I can think of one and trying to think if it's, uh, something I can share. Um, And I'll, I'll give you an example and I'll, I'll keep it generic is a video analytics solution. And, um, uh, in, in that situation, they were looking to implement video analytics and run that through an edge solution, [00:14:00] um, where there may be cameras that are going to be capturing video of, uh, public places.
[00:14:09] And, um, in doing so the cameras are going to sit on a, uh, a public network or at least a, a network. That's not part of, um, a private network. So you have to worry about security and access from those cameras to the internal network. But then when the video is being processed locally on the edge device, You also need to be able to give access to the end customer that's going to be using this device.
[00:14:37] So think about, uh, it could be, in a public space, in a campus arena, whatever it may be. And, and so that end customer that is actually implementing it also wants to be able to view the video and view the, the analytics that are coming from that device. So you have to look at again, the firewall between.
[00:14:59] Let's [00:15:00] say the customer's network and the network provider's network, as well as, um, any of the suppliers that might come in to be doing services to support that solution. And so that's, that's where you have to really look at what is the architecture that's being put in place, because when you start looking at the data flows of not just the video, but also support services for, uh, people that are going to be maintaining the systems, as well as.
[00:15:29] Um, any other data flows that architecture could change based on where the data needs to flow and as well as the latencies and, uh, the bandwidth capabilities, all of that comes into play. So
[00:15:44] Matt: [00:15:44] super interesting. So let's, let's do this. Let's do this. It'd be kinda fun. I think let's, um, let's, let's design a hypothetical. System. All right. Not one, that's tied to anybody in particular. Okay. So let's say we have a campus, uh, UC Berkeley's right down the street from me. And let's imagine that, um, because [00:16:00] of our concerns for safety and other things, we want to place a bunch of cameras around the campus and, uh, utilize the data from that campus.
[00:16:09] And we have multiple vendors that might do image analysis or security, and maybe I need the, the campus police to get access to it. And so I envision this, so where, where do I even start?
[00:16:18] Rodney: [00:16:18] Yeah. So you're going to start by having a conversation with the customer about what are their goals, what are they looking to achieve? And, you're going to have to take a look at what are the areas at that campus that they want to monitor. So if you're saying, at the campus, it's, how many locations is it going to be?
[00:16:36] The dining facility? Is it going to be, in one of the auditoriums or, or wherever it may be, that's that's part of it, then you also have to,
[00:16:45] Matt: [00:16:45] We've, let's just make this up. We've identified a hundred locations. Some of them indoor, some of them outdoor, but all public where there's no expectation of privacy. Um, some of them may have a wifi. Many don't some are outside. So I'm going to be on the top of a pole, maybe [00:17:00] mounted to a tree or the side of a building, like now, what do we do?
[00:17:04]Rodney: [00:17:04] Um, so usually it involves bringing in a partner who does the video analytics to you do a site evaluation? Uh, usually we'll have a conversation initially on the phone, but, eventually it turns into a site evaluation to look at those those hundred locations. And then what type of, um, video or what type of services are you looking?
[00:17:24] What type of analytics are you looking for? Um, and then from there, we'd work with the partner on establishing, based on the, the customer input for what types of video and what type of analytics they're looking at doing. How much compute resources are going to be required. What types of networks are going to be required?
[00:17:41] So if you're using wifi, if you're using wired, or if you're using a fi uh, four or 5G type cellular connection, that's going to have dependencies as well as, any of the, um, the security measures that need to be put in place. As I mentioned before for firewalls, and you start building a solution.
[00:17:58] And you start with a, an initial rough [00:18:00] draft design. That's going to say, okay. Based on those a hundred locations, based on the type of analytics and data collection you're going to be doing, you're going to need X amount of compute X amount of storage. Um, and to be able to process that at the edge to give the customer what they're looking for, whether that's people counting, whether it's geo-fencing, whether it's, um, any, any other video analytics you want to do, um, Going from there, you then refine that design by looking at having those discussions around the data flows, who needs access to which components.
[00:18:36] Again, if you get back to, vendor doing a support of the system versus a network provider versus the client, the, the university themselves who needs access to what and where, and, and then ultimately from a, a. Uh, relative real time, um, capability, what are the KPIs that [00:19:00] they need to be able to make that a, a functional system and, and you put that all into the solution.
[00:19:04] So picking and choosing the types of infrastructure picking the type of networks is all going to be dependent upon. Every all of those components, especially the KPIs, if, if they don't care about having video in real time or real time, but maybe they want it near real time and they're going to store some offsite.
[00:19:22] Then that's one thing. But if they're gonna store everything onsite at the edge that changes the dynamic of the type of infrastructure and the type of networks you need to be putting in place.
[00:19:30] Matt: [00:19:30] Yeah. So let's, let's, uh, let's, double-click on some of that. So in this hypothetical instance where I'm responsible for deploying these cameras at UC Berkeley, um, you mentioned private versus public networks. Um, how should I even think about what? So what is a private network versus a public network and how should I be thinking about that?
[00:19:48] Um, as, uh, as, as someone who's looking to install video analytics,
[00:19:52] Rodney: [00:19:52] Sure. Public would be, anything that might be internet connected internet round-table. Um, private, what is going to be the [00:20:00] university's own network, where you have to be on the university's network to be able to gain access to it. So you couldn't access it through the internet unless you were going through let's say a VPN or something like that.
[00:20:11] Um, and I'm sorry, the other half of your question, right?
[00:20:16] Matt: [00:20:16] I was relying on you to remember that. Um, I was asking should I think about, um, uh, private versus public and where I really wanted to go with this? Rodney is, um, I wanted, when I think of public networks, I think of Verizon at and T Comcast and so on. And I think about a university, a university environment.
[00:20:33]I imagine that the university. Network has grown organically and probably was not built to accommodate real-time video analytics. So is there, are there other suppliers that can just give me a network that has security, whether it's public or private, but can actually like build my network for me for this video analytics, are there companies that do that and, and how would I, who would they be and how would I even [00:21:00] think about involving that versus extending my own network into this space?
[00:21:04] Rodney: [00:21:04] Yeah, absolutely. And, you mentioned it, with the, uh, the service providers, the, the at and T Verizon, uh, other companies that do the network services type of, um, services and those networks. Can be, a private network for the customer, whether they're, building out a wireless network, you see this in public spaces all the time where, you go into your local coffee shop and they have, free wifi, that's more of a public approach, but then there's the, you go into the, uh, the university bookstore or are you going a better, better example is going to be, uh, the university classrooms where they're going to have their own. Wifi network. That's separate. You can't access that from, uh, from outside.
[00:21:49] And so to be able to build that, you're going to be working with a network provider on, on those networks and, and what what's required for, um, [00:22:00] for the amount of services that are required, how many students are going to be using it, how many faculty are gonna be using it. Um, and, and building it out there now in terms of the video analytics and what you want to do there, um, It's it really comes down to the, the requirements of the application.
[00:22:18] Matt: [00:22:18] Cool. Let's, let's say that, that, that I'm really ambitious. So I want a system that can do real-time analysis. I want it to be extensible. Um, so that I don't, I don't really know what I'm going to need, but I want to be able to add capabilities to my network, my, my analysis pretty quickly. Um, I don't want to maintain the network.
[00:22:37] Um, and let's say that it actually has to accommodate crazy stuff. Like I want. Uh, let some cameras mounted on, on, uh, some of the security, uh, uh, uh, golf carts that, that we drive around at night to keep students safe. And so there's, there's gotta be some mobility. So I'm gonna hire somebody like at and T or somebody like that to, to provide me with a network that has all the security and all the, the mobile [00:23:00] handoff and all that clever stuff.
[00:23:01] And then. Like w what happens. And I realized that that mobile edge computing, uh, uh, or multi-access edge computing, depending on, on, uh, how, how long you've been familiar with the ETSI standards, um, how that fits in. So extreme to the audience, the, the different, big building blocks, and then how mobile edge computing can help me solve some of these problems.
[00:23:26] Rodney: [00:23:26] Sure. I'm going to break that into three parts. Cause you talked about ETSI MEC, et cetera. You're asking about the golf cart scenario. Um, it, it really depends on the network provider on, whether they want to follow. At T Mac or if they want to do their own kind of solution for, for, for Mac, for multi-access edge compute.
[00:23:48] And, um, I'm not gonna really get into the debate over whether it should be Etsy Mac, or if, uh, if, uh, if they want to do their own. Cause it really, really comes down to being able to provide a service to the [00:24:00] customer. And I, under the covers, I don't think the customer really cares what, um, architecture or what standards was followed as long as it provides the same experience to the customer.
[00:24:10] So then if we get into your experience with the golf cart and put a camera there, um, you have to look at how do you get the video streamed wirelessly from the golf cart to some fixed network that can actually pick that up. So you could do wifi, you could do a cellular network. Um, if you're close enough, you could do Bluetooth or you could do, um, any other of the wireless technologies, um, in that situation, I'd probably, um, Be looking at probably a cellular network simply because you're probably outdoors not going to be doing that many, um, wireless access points. You may, maybe it depends, the campuses do have that. Um, so I would say it would either be wifi or cellular for, for that connectivity, but then once you get beyond that, you're going to come into a, a fairly fixed network.
[00:24:59] That's gonna. [00:25:00] Um, aggregate that data, the video traffic from those cameras, and then you're going to send it to an edge device for processing that edge device is going to sit fairly close to the edge of that fixed network to be able to aggregate and process that data. Especially if you're looking for, um, any artifacts within the video, maybe you're looking for security purposes and you want to identify specific.
[00:25:23] Characteristics of that video, it could be the color of the clothing could be the, um, the shape of the object, whatever it is, if you want to keep track of all your golf carts, you could look at the shape and identify and put it through a, uh, a machine learning algorithm to be able to identify, Hey, that's what a golf cart looks like versus, a car or a truck, et cetera.
[00:25:43]Um, but getting back to the architecture is within that edge device, it's going to process and store the video that's been analyzed and be able to present that to whoever's viewing it. And it can do a couple of things in real time or near real time. [00:26:00] It can trigger based on whatever you've programmed it to do.
[00:26:05] To send alerts. It could, um, notify authorities if you wanted to do something from a security standpoint, or it could just go into a, a, a, a system for reporting that sends a report to somebody to say, Hey, um, this is what we've noticed in your video, wanting to make you aware of what's going on. So a lot, lot of different things that can be done there.
[00:26:26]Matt: [00:26:26] And you mentioned the, um, macro not Mac and, and, and all that. So help, help, help us understand, uh, what is the purpose of. A multi-access edge computing. Like what, what is that? That's a lot of words. What, what does it actually mean? How do I use it and why should I care if it's a proprietary mechanism or a standard coming out of ETSI or not?
[00:26:54] Rodney: [00:26:54] I think for, for the standards coming out of ETSI, it gives [00:27:00] people that are going to be developing for MEC a way to make sure you have standard interfaces, APIs, you know, that that gives you the capability to have, uh, some commonality across platforms, across applications and the way things are implemented.
[00:27:16] Matt: [00:27:16] If I'm a company that does video analytics, I do people counting. Um, and I can, I can build a software interface to the MEC API, um, and be fairly confident that everybody that's implemented MEC, um, would have a pretty straightforward integration with my technology. Is that correct?
[00:27:37] Rodney: [00:27:37] for the most part. And I would probably defer one of my colleagues actually sits on the chair for ETSI MEC. Um, I would always defer to him to, uh, to answer those questions, but, um, in, in the basic sense, yes, I think that that's what it allows for now. There's always going to be each. Vendor or each provider's solution or implementation [00:28:00] of, of MEC.
[00:28:01] So those, even those though those interfaces are defined by ETSI MEC. You may find that there are some variations and that's why I'm um,
[00:28:09] Matt: [00:28:09] Yeah. And actually I wasn't, I wasn't, it wasn't trying to take us into the weeds there. What I was trying to get at was, um, this is pretty powerful idea that there is an interface that's programmatic between the radio network. The private network or even the public network, if that's the application that would allow me as a video analytics company to really specialize in video analytics and not have to know anything about the wireless network, really just to be able to talk to some interface to say, okay, send me a video feed and let me.
[00:28:43] Do something with it and alert somebody or whatever the action is. And, and I think having those interfaces defined, um, whether it's a public standard or a private standard or, or a proprietary extension to a public I, whatever it is, as long as, as, as most of the code I've already written [00:29:00] can still reapply.
[00:29:00] It feels very powerful and it feels game changing. Um, but I want your perspective on that.
[00:29:05]Rodney: [00:29:05] In the, in the essence of, pretty much applications, they, for the most part should be agnostic to the hardware or to the, to the architecture. And in that sense, It, it comes down to the ability of being able to, if there is an API that has to be written to interface, usually it's something around a control mechanism or a monitoring management mechanism.
[00:29:27] And those are usually within the application, some of the APIs that might have to conform to an ETSI MEC, um, But in the true sense for most applications, you should be able to run many applications agnostic to the architecture or to the standards, in my opinion.
[00:29:46] Matt: [00:29:46] And, and, and that's super powerful as a solution architect because you have lots of vendors that you can pick and choose from to complete entire solution for the end customer.
[00:29:58] Rodney: [00:29:58] Absolutely. And [00:30:00] I wanted to get back because you were asking why MEC right. And, and um, a couple of reasons it's from a solution provider or from a network provider, we, cause you have the network edge and then you have the customer edge. So you have two areas where you can place, um, place the devices.
[00:30:18] It could, if you go into the network provider edge, um, they're going to be placing. Devices within their own network. And when we say MEC multi-access, it could be wired, could be wireless, could be whatever connectivity that's where the multi-access comes into play. A lot of it is tied around cellular networks because today a lot of companies do already have wifi networks, so they could implement edge solutions using their own existing networks.
[00:30:47] But with MEC. It's it's really coming from the network provider as a solution for their customers to bridge the gap between the cellular network and the enterprise network. [00:31:00] So the customer's enterprise network. So if you think about today in cellular networks, all the data that's coming across the cellular network.
[00:31:08] If, if I take a, a campus or an arena and they have cellular services for customers that come into that arena, All the cellular traffic today without MEC doesn't allow any of that data to traverse into that customer's enterprise network
[00:31:25] Matt: [00:31:25] that's right. It might go through Dallas.
[00:31:28]Rodney: [00:31:28] It stays on the, the, the service providers cellular network with MEC that gives you the bridge and gives you the ability to traverse and hand off any certain traffic.
[00:31:40] It has to be determined by, the customer and the network provider on what that traffic is going to be. Let's use the video analytics example. If, if I have a camera that's connected over 4 or 5G to the cellular network, how do I get that onto the enterprise network? You have to have a bridge.
[00:31:55] And that's where the mec device comes
[00:31:57] Matt: [00:31:57] inTo play
[00:31:57] Yeah. So an example of the stadium, I've got a bunch of, [00:32:00] uh, wireless cellular wireless access points. So I want my fans to have great cellular service, uh, in my stadium, but I'm also gonna use that network for my security cameras and the security camera data. I don't want it going on the. The public network.
[00:32:17] I want it to stay in my facility. I want it routed into my, my, uh, video analytics servers on the facility or nearby. And I want my onsite security people to have access to that. Um, and, and MEC and the surrounding technologies are what makes that happen? Um, that's, that's really powerful. So let's switch a little.
[00:32:37] So Hewlett Packard enterprise. So Hewlett Packard enterprise, uh, it's been around a long time. Uh, actually Hewlett Packard enterprise. Hasn't been around that long, but he'll Packers to company around long and, and Hewlett Packard enterprise is among other things in the business of making servers. Um, and as I understand it, yeah.
[00:32:53] That Hewlett-Packard recognized that there was a unique set of needs in edge [00:33:00] environments. And you have a whole product line on, on that edge servers. Can you tell us a little bit more about what, where that thinking came from and what solutions uh, your company has developed to address the unique needs?
[00:33:14] Rodney: [00:33:14] Absolutely. And when we talk about our. Edge products. It's our edge line product line. Um, it really started when we had, if you look at the different industries and my background is telco, but let's go outside and look at other industries, you get into manufacturing, you get into oil and gas, et cetera.
[00:33:31] And a lot of devices that are generating data typically were analog have moved to the digital age, so you have, Programmable logic controllers. You have smart motors, smart, no pressure sensors, whatever it may be. That's in the industry. And you now have the need to be able to take, collect and process that data at the edge and create a real time loop that, for example, if I have a motor that's being monitored, and let's say it's [00:34:00] in an oil and gas industry, and those motors are expensive, very expensive.
[00:34:03] And they're out on an oil rig that are very far away from the mainland. The oil and gas customer wants to be able to look at those devices and understand what's happening with, with their, uh, their products. And so you take an edge device. That's been developed for those types of environments. So you can't necessarily take a data center server and stick it out on an oil rig.
[00:34:27]It's a harsh environment. Um, the, the space, the cooling, all of that that's required for a data center server may not be applicable in, in that type of environment. So we looked at it to say, how do we develop a, a system that's hardened that can withstand the environmental impacts? That's a small form factor.
[00:34:47] And is, is, capable of operating within the customer's requirements for those types of locations. And so you, you do that by developing a new whole new product set, and that's where edge line came in. [00:35:00] Um, so we have, um, A few products that fit different needs everything from entry-level compute all the way up to a high end, uh, compute that leverage a lot of the latest processors available in the regular data center servers.
[00:35:13] And so you can run the applications that you typically run in the data center, but now you can run them at the edge. And that's where our edge line product line really fits.
[00:35:22] Matt: [00:35:22] So when I, when I think of, um, data centers and add in servers in general, one of the trends in, in data centers, environments is the amount of processing capability, storage capacity. GPU's all these things, um, that getting very, very dense. You can be able to get a lot of compute in a small form factor.
[00:35:43] And one of the trends that's, that's driving, it's requiring very sophisticated, critical infrastructure for some of these very high end. Uh, servers, like the kinds that run in and, Facebook and Amazon and Google and something like that. And these critical environments, they are, [00:36:00] they've got very sophisticated cooling systems, airflow systems, all these things, um, which at the bottom of an oil rig might be inappropriate.
[00:36:09] And so is it, is it the case that, that the Edgeline servers, um, uh, have, have contemplated not being in these very. Um, sophisticated data center environments, but out in locations and environments where it's not nearly as, uh, predictable, um, is that the case or do I need a mini data center environment? Um, in, in my next to my, my oil rig.
[00:36:36] Rodney: [00:36:36] So it, it depends. And um, if we're looking at, let's say they'll take a hundred. Systems that are operating to maintain all the pumps and systems at an oil rig. And I want to collect all of the sensor data from each of those pumps. You might place an edge device within relative proximity to where those pumps exist, but then you may also have a [00:37:00] mini data center sitting elsewhere where you're going to do the edge processing on that one device connected to the pumps.
[00:37:07] And it's going to send real-time data back to be able to provide, um, Either predictive analytics or maybe a pump fails and it's able to redirect to it, to another pump, but then you're going to have data that's still coming off of there that needs to go back for more longer-term analytics or reporting capabilities that might go into more of a data center function.
[00:37:28] So you, in the oil rig case, or maybe in a, in a manufacturing facility, you might have a small, uh, data center pod or something that would be also be tied to that in the overall solution.
[00:37:41] Matt: [00:37:41] Yeah. Yeah, it does. I think that, that, um, that lines up, uh, really sensibly with some of the conversations that, that we have in the industry around, uh, it's. It's well, in some ways I talk about five years from now. We might not actually be talking about edge anymore. It's all going [00:38:00] to be part of the internet.
[00:38:00] And really there is a continuum from the device. All the way back to some centralized cloud that might be in Seattle or Washington or something. Um, and that really our challenge as, as, uh, practitioners in this world is to build the tooling and distribute these systems so that you can literally say okay, I've got, I've got, uh, three services.
[00:38:25] I need to process this analytics data. Some of it I have to have right next to the device. Some of it, um, either because the data's too expensive to transport or too, too proprietary and I need secure or whatever reason. Um, and then I have a, something, maybe it's a little bit closer, but it's a data center environment.
[00:38:43] Um, but it's, it's a constrained environment. So I can't, I can't scale up infinitely. It's just kinda what is what it is, it's at full capacity. And then I might have some cloud resources and some of those cloud of new sources might be in, Service provider edge locations. So part of the infrastructure side of the last mile [00:39:00] network, and some of them might be in just traditional cloud data centers, and I'm actually computing across that entire spectrum and delivering, uh, a service that has optimal placement of workloads and, uh, it was using each type of facility for what it's best at.
[00:39:21] Rodney: [00:39:21] And I guess I missed the question.
[00:39:23] Matt: [00:39:23] Oh, there wasn't really a question. I was. I was,
[00:39:27] Rodney: [00:39:27] No, very good points.
[00:39:28] Matt: [00:39:28] do you agree, dude? That's the
[00:39:30] Rodney: [00:39:30] Yeah. Yeah, no, I think, I think you hit a lot of the good points. Um, because really, when we look at what's happening across the world is, take the internet. It is a conglomeration of networks, it started off, as just a couple of universities starting to communicate and it grew, exponentially.
[00:39:50] And when you took, when you look at. I mean what we're doing today with just video conferencing, I don't know where [00:40:00] things are connecting. But somehow it's happening. And so same thing with the edge and the cloud and the data center and the, other, mini clouds or whatever you might have in place.
[00:40:12] It is going to be part of an overall solution that needs to, have, uh, You need to look at what is really the application that's going to be running or set of applications or platforms and what are the requirements for each? So some things will operate in the data center. Some will operate in the cloud and some will be at the edge.
[00:40:34] And you're going to use all of those as a tool set, to be able to develop the overall solution. So I think you hit on a lot of good points.
[00:40:42] Matt: [00:40:42] One of, uh, uh, the things that I'm continually fascinated by is all the activities. He's in the ecosystem that are contributing to a much more accelerated realization of value. [00:41:00] Um, the, the, the sort of business school ways, time to value, right? And that's like how I have this idea. I want to automate my factory, or I want to, automate my, uh, my, my precision agriculture.
[00:41:10] Um, and. Yeah. Is that a decade long problem or is that a quarter long problem? Um, and I think five years ago it was probably a decade. See, it felt like a decade long problem. Uh, and now it's feeling more like a, like a quarter problem. And I think part of that is what we talked about earlier. It's like open source, right?
[00:41:27] There's lots of. Pieces that are being built for me that have been built for me, that I can, I can leverage, I can stand on the shoulders of those giants and implement a lot of the solutions. And my, my vendor suppliers can implement a lot of those solutions. And that gives me faster time to market and flexibility and the knowledge that I can switch vendors potentially if that needs, if that's an important criteria for me.
[00:41:47] Um, so I've got all of those components, but there's another component that I think is really interesting and that is the business models. And I think there's two really big contributors to, to, uh, substantial shift in business while there might be [00:42:00] more. But these are the two that I think about one of them is, is the economics of shared infrastructure.
[00:42:04] And what I mean by that is, just take the cloud example. The fact that I can share a cloud computer with 10 other people or a hundred other people, um, and only have to pay for that fraction that I use, uh, means that fundamentally changes the economics of, of how I, uh, deploy compute. Um, because not, not only dimension that because I'm sharing the infrastructure, but there's another piece too, which is I don't actually have to buy the infrastructure.
[00:42:30] And those of us who are in the business of building and deploying infrastructure, uh, Ourselves, meaning like vapor IO. We, we build our own data centers. We lease the land, we put the building on it. We trench the fiber. We do all these things like we're used to building this infrastructure and then selling it or leasing it in pieces to our customers.
[00:42:51] But servers outside of the cloud environment have never really been purchased that way. And, uh, you mentioned before the interview that Hewlett [00:43:00] Packard enterprise is on a path to deliver everything it does as a service. And how do you deliver computers as like physical computers as a service? How does that done?
[00:43:13]Rodney: [00:43:13] So it becomes a kind of a model that we put together that allows customers to be able to consume only the resources they need and by doing so we work with the customer on developing a, a plan that looks at what is their immediate need for servers. So let's say they need 10 servers. And we also look out into the future for, for a forecast to say, How many servers do you think, you're going to need over the course and maybe looking, every quarter or whatever it is, and they're going to pay for an upfront cost of what kind of infrastructure they need.
[00:43:52] And we can put it into a, a, a monthly model to just to consume just like you would the cloud, you, you pay your bill [00:44:00] monthly. Um, but then we're going to monitor that their usage. And we're not monitoring the, the actual data that they put on there. We're monitoring the server performance, the server capacity, et cetera.
[00:44:10] Matt: [00:44:10] The value that I'm getting out of it.
[00:44:12] Rodney: [00:44:12] Yeah, exactly. And we're looking at what is the utilization of that infrastructure that, that, that customer is consuming. If they're consuming 80% of that infrastructure day one, we're also going to implement additional capacity upfront to be able to give them the capability to grow. Beyond that, 10 servers.
[00:44:35] So instead of 10 servers, we might put 15 servers in up front, right? So they have an extra window to be able to grow immediately. And as soon as they start using it, then, then they're going to pay for it. That's part of the GreenLake model. But the other part is that. Capability to be able to consume it on a, on a monthly, pay as you go type charge.
[00:44:56] So the, the whole GreenLake model is around, um, [00:45:00] what we call flexible capacity, which is what I described there, where it's, a bit upfront and then use as you grow. And if, if you don't like, let's say, um, think of, think of the. Consumers where it's based on, um, the retail market and things fluctuate from, the holidays gets really busy and then, you have the lulls, right?
[00:45:20] So in those models, a GreenLake model is, or a flex capacity model is a really good fit because they're going to pay for what they need during those really, really, um, busy times. And then when they're in a lull, they can scale it back for customers.
[00:45:35] Matt: [00:45:35] So it's, it's, it, it is really almost like how I pay for a cloud provider. Like I can scale up and scale down to S to some extent, if you've only got 15 servers or deployed this as high as I can go, but I, but what I pay to Hewlett Packard can scale up and down based on the seasonality of my business.
[00:45:51] Is that what I'm hearing?
[00:45:52] Rodney: [00:45:52] To some degree within certain parameters, but, um, other customers that aren't, so fluctual, they, they don't [00:46:00] fluctuate so much. Um, W, if you take a new platform, um, when I was doing network planning, sometimes you just don't know how much capacity you're gonna need upfront. Um, the launch of the iPhone, um,
[00:46:13] Matt: [00:46:13] yeah.
[00:46:14] Rodney: [00:46:14] caused a AT&T just, numerous amounts of network expansion because of the popularity.
[00:46:20] And when you start a new platform and you just don't know what the capacity is going to be. You have to start somewhere with Greenlake. That gives our customers the ability for us as a vendor to supply them infrastructure on an as needed basis. So we're going to give them X amount of capacity upfront and.
[00:46:37] If they start consuming more, we're going to monitor it based on the performance and the KPIs of that infrastructure. And we can bring more infrastructure in. It's not just infrastructure. It's the services that we provide at Hewlett Packard enterprise.
[00:46:50] Matt: [00:46:50] Rodney Richter on demand.
[00:46:52] Rodney: [00:46:52] Yes. Um, no, we have a whole point next services team that can do, support services as well as, um, uh, services [00:47:00] for deploying the infrastructure and, um, monitoring and management, et cetera.
[00:47:04]Um, that's all part of that Greenlake model to be able to do that as a service. And, we are the edge to cloud company. That's offering everything as a service and that's what HPE is
[00:47:14] Matt: [00:47:14] That's that's, that's really, that's really neat. And the other thing that's, I think pretty cool about that is, um, it feels to me like the customer relationship is profoundly different because it sounds like Hewlett Packard is taking some risks. Like you're taking some financial risks. In, and I am too, I've got some upfront fees, but you're taking financial as can forward deploying capacity that you anticipate me using.
[00:47:35] Um, but not necessarily on the timescale, that would be ideal for your quarterly financials. Um, and so we're sharing an outcome, which is, uh, uh, a really nice way to have a relationship with a vendor, I think. Is, are you finding that that's, um, uh, something that customers are getting excited about?
[00:47:54] Rodney: [00:47:54] So I can't speak to the financial aspects. I'll leave that to my leadership, but I think customers are [00:48:00] excited that, it's a, it's a new way of doing business. And it gives them new capabilities that they didn't have before, before. It was always pretty much a capital purchase. I need X amount of infrastructure and Oh, by the way, my budget says, I have, this amount of capital and yeah, I'm going to have to spread that out over the next couple of years.
[00:48:21]Now
[00:48:22] Matt: [00:48:22] And put it on my, on my balance sheet and appreciate it and all that other stuff.
[00:48:26]Rodney: [00:48:26] So now they can look at it and say, rather than come up with a very large sum of money up front, they could pay and turn it from a CapEx to an OPEX model pay per month for. The infrastructure that they need right now, and then pay a little bit more if they use more, pay a little bit less, if they use less.
[00:48:45] And that's, that's the beauty of it is it gives them that flexibility and really gives them more of a cloud like experience. And we can do this on-prem and we can do it off prem. So if a customer wants to put it into like a, a vapor IO data center, or, a hosted data [00:49:00] center, we can put our infrastructure there and host it for the customer.
[00:49:03] Matt: [00:49:03] Yeah, that's, that's, uh, that's really exciting. I think, I think it's, uh, it's, it is a profound difference, um, because it is, it is, uh, the whole world is moving towards as, as a service and, I think about. I, I don't want to own any servers. I want the people that know how to, how to maintain and, and, and replace and do all it to own servers.
[00:49:24] And so it, it seems like a really good division of labor, back to this like core business principle, which is I want to focus on the things that are most differentiated and honestly competing against Hewlett Packard, so to speak in. Racking and stacking and operating and financing servers just doesn't make any sense.
[00:49:43] So I'd rather be a partner of yours. And when I say competing, I don't really mean competing because I might buy them from you anyway. But you know what I'm saying? It's, it's that is really an expertise that you're building and you're driving all the, the efficiency into that. Um, so you're going to be able to operate, maintain, and supply servers a lot better than I [00:50:00] am.
[00:50:00] I think, I think that's, that's how I would look at it.
[00:50:03] Rodney: [00:50:03] Yeah. Then that's what we do as a business, right? That's, that's our bread and butter is servers. Server storage, networking, and and we've also expanded out to, um, look, we've acquired a few additional companies over the last couple of years to really expand our portfolio and really bring a, a solution level to the customer.
[00:50:23] And, part of GreenLake is not just infrastructure. It is, it could be infrastructure as a service. It could be platform as a service. Um, It really depends on what the customer is looking for. And, um, we also partner with a lot of different companies to be able to bring in the different applications and, and, uh, the total solution to the customer.
[00:50:43] Matt: [00:50:43] Great now let's, I'm going, gonna ask you to take off your, your HPE hat and put on your member of the ecosystem, LFEdge participant hat, and, um, you, you think about the, the. [00:51:00] The dynamic between the cloud providers and the telcos. And this is really interesting to me, um, where they seem to really need each other.
[00:51:09] Um, but they also, um, seem wary of, of potentially competing with each other. And I'm wondering if you have a perspective on, on, on that dynamic and, and w how you think it might net out.
[00:51:22]Rodney: [00:51:22] Um, I, I agree with you that, they, they need each other, um, in that, with a lot of the cloud capabilities that are out there. I, I think, uh, most of the service providers are leveraging. All different types of technology, including cloud. And uh, so th that's one aspect of it.
[00:51:41] So they're, they're already using cloud and, um, and many different types of the big cloud providers, but at the same time, you're seeing the cloud providers come in and provide various different services that do compete with those service providers. Take edge, for example, um, and I don't know if we want to go into [00:52:00] specifics, but, uh, Google cloud platform, um, Uh, AWS outpost, those types of edge services.
[00:52:08] I shouldn't say Google cloud platform itself. Google anthros, I think is what it is. But, um, either, either way, bringing the, the cloud services on prem. And, um, in some ways providing solutions that a service provider might provide, like multi-access edge compute, for example, that could be one area where they may compete or, um, when you start looking at any of the applications that typically a service provider, a telco provides, um, voicemail services, um, any of the data, um, Processing capabilities that, that you find today from those service providers, some of the cloud providers are moving in those directions.
[00:52:53] And yeah, I think, I think there is some competition there. And, um, so it is, it is a give and take [00:53:00] between the two.
[00:53:01] Matt: [00:53:01] Yeah, it feels overall healthy to me. Um, yeah. Cause the other, the other thing is, you think about outposts, you think on the surface it might be competitive, but I could become an outpost reseller. Because I've got the field sales team, uh, I going to build the wireless network, potentially that services, that, that wherever those outposts are going to go, um, I, I can deliver potentially an end to end solution without post or with Hewlett Packard enterprise or whatever.
[00:53:25] So that's, it is, that is, that is interesting. So on the edge computing front, um, what are you seeing as the types use cases that are getting traction today? Like what, what does, what, what w. Where's the money seemed to be going. And how is that changing?
[00:53:44] Rodney: [00:53:44] Um, in some ways I'm thinking, cause I, I work specifically in telco and um, at least within the service providers, the network edge is, is a lot. And then the work that they're doing to offer MEC [00:54:00] to their, their customers. But as I've talked with. We'll say the end clients, as I've gone to some conferences and, and met with various different industry, um, clients, I found that everybody's looking for, ways to optimize their business, that, as they invest in edge it's it's, what can they do different?
[00:54:21] How do they make it more efficient? And I think a lot of it manufacturing is, is big. Um, oil and gas has big. Um, we're starting to see it in retail and hospitality. Um, looking at things and ways to automate at the edge to be able to, do things that, they've been done before, but haven't been done as quickly or as efficiently.
[00:54:46] And um, a lot of those areas, uh, a big one, um, racing industry is actually using edge computing to improve. How they look at the, um, the [00:55:00] cars that are being, uh, developed and, and tweaked for the racetrack. And they use sensors on the cars to be able to go around the track and collect the data and then refine how they're, they're tweaking the cars for the race.
[00:55:12] So
[00:55:13]Matt: [00:55:13] An, an F1 car is just a, a factory for delivering speed around a track. And you can see how those applications that are developed in that very, very bespoke and uniquely competitive environment will make their way into traditional manufacturing. That's that's a neat that's. I'm gonna have to look into that more.
[00:55:33] That's a, that's a, that's a neat application of edge computing is in, is in racing.
[00:55:38] Rodney: [00:55:38] It is, and it's exciting, exciting sports, so great to watch and great to be involved with.
[00:55:44] Matt: [00:55:44] Yeah. And where, where do you see edge computing going over the next 18 months?
[00:55:50] Rodney: [00:55:50] Over the next 18 months, I think, um, we're going to see the evolution of, of the network edge a lot more. Um, I think [00:56:00] from, uh, at least in the open source community, a lot of applications and platforms being developed that are going to help, um, Move things forward, uh, especially from evolving from some legacy platforms and seeing the data move from the data center out to the edge.
[00:56:19] Um, I know we, you were talking on your state of the edge previously about the amount of, uh, infrastructure and data. That's going to move to the edge over the, what was it? I'm looking back at your notes. Uh, server storage infrastructure needed to support edge $700 billion and 200,000 mil megawatts of compute power needed.
[00:56:41] What was that timeframe?
[00:56:43] Matt: [00:56:43] 10 years.
[00:56:44] Rodney: [00:56:44] 10 years. Okay. So we're talking 18 months. Um, it's, it's gonna probably be, uh, a big chunk because we've already been on the edge journey for some time. HPE started the edge journey, um, probably one of the first companies to do and, um, so we've been in the [00:57:00] edge business for quite a few years now and I just see the next 18 months just starting to explode really.
[00:57:06]Customers that are going to, um, either implement it. At their customer sites or they're going to be using at the network edge.
[00:57:14] Matt: [00:57:14] Yeah. And we're recording this in September, 2020 in the middle of the COVID lockdown and it, it certainly. My, my conversations have indicated that a lot of these, um, uh, edge computing automation capabilities are the spending on that as being accelerated. Are you seeing the same thing?
[00:57:34] Rodney: [00:57:34] Um, to some degree, COVID did, uh, cause some pauses in some of the projects, at least, um, the ones that I've been involved with, but, um, No, a lot of that has to do with, when you consider
[00:57:47] Matt: [00:57:47] just general business disruption.
[00:57:49] Rodney: [00:57:49] just, yeah, exactly. You, you don't have people going out like they did before, especially in the public spaces, but, um, but now there is, we're starting to see [00:58:00] customers spend more for how we operate in a, in a new COVID environment where, um, Yeah, you want to be able to provide the resources to your customers in a different fashion where they're no longer necessarily onsite, especially, um, security is a big one, um, you now have facilities that aren't as populated as before.
[00:58:23]You may still need to monitor those facilities, um, in real time to be able to know what's going on, um, especially with the climate and the, the, uh, What's happening around the nation. We want to be able to make sure our facilities are secure. Um, think, think about a facility that's been sitting for a while.
[00:58:43] It has nobody in it. The, probably the air conditioning systems or somethings, but that's been shut down is the air quality in there. Good for people to return to, to the office. Um, do you need to do any monitoring there? And once people start coming back, do you need to monitor? I got a one that's, that's actually a good use case that, [00:59:00] um, We're hearing about, and that's, you go to the hospital or you go to your doctor and they measure your temperature before you walk in, and it's somebody standing there waiting.
[00:59:09]What if you could automate that?
[00:59:11] Matt: [00:59:11] You should be able to write just camera's consents, infrared,
[00:59:15] Rodney: [00:59:15] Yeah, exactly. So you have a camera that takes your temperature, that, um, there's mask detection to make sure people are wearing masks. There's, um, there's all kinds of things that can be done with video analytics to be able to do that amp. And it's not just being able to sense it through a video it's being able to then what do you do with
[00:59:32] Matt: [00:59:32] actuate something.
[00:59:33] Rodney: [00:59:33] Actuate? Yeah um, touchless, um, Access entry to a building. Um, I was working with one company where that's what they do. They do video analytics and their whole building is automated to where they just do facial recognition. And it,
[00:59:50] Matt: [00:59:50] So this, so my face is my ID card into the building.
[00:59:53] Rodney: [00:59:53] exactly. And if you have a mask on, they can even detect whether or not you're who you are based.
[00:59:59]There's [01:00:00] certain characteristics that they look for, but, um, it's, it's very interesting, technology that they have.
[01:00:05]Matt: [01:00:05] I was told by a video analytics person this was years ago, but that they actually can detect tech, detect people individually, by the way, they walk by their gate. So it's not, you can do facial recognition, but you can also do like ambulatory recognition, which just seems completely fascinating to me.
[01:00:21] Like we're running a completely new world.
[01:00:23] Rodney: [01:00:23] That that is amazing. Yes.
[01:00:24] Matt: [01:00:24] Yeah. Rodney, I really appreciate you joining us. This has been a fascinating conversation. Uh, we didn't get to half of what I was hoping to talk about, but where we've, we've filled the hour, um, if people want to get ahold of you online, what's the best way to do that.
[01:00:38] Rodney: [01:00:38] Um, they can reach me via my email address, Rodney dot richter at hpe dot com. Um, that's probably the easiest way. Um, I'm also on LinkedIn um, feel free to reach out. And Matt, thanks for hosting and having me on the show.
[01:00:52] Matt: [01:00:52] Yeah, you bet, Rodney. I really appreciate it. Enjoy the rest of your day,
[01:00:55] Rodney: [01:00:55] All right. You too. Take care.