Hewlett Packard Enterprise Company (HPE) Earnings Call Transcript & Summary

Thank you, Simon and Eric, for the welcome remarks. My name is Christian Renaud. I'm the Research Director for the Internet of Things channel here at 451 Research, a division of S&P Global Market Intelligence, and welcome to the Edge of Tomorrow session – 5 Ways To Capitalize on the Edge Opportunity. This is a live session. So in keeping with 2020, please bear with us for any abnormalities that happen as far as broadband or audio or video cutting out or cat fights or my dog snoring or anything like that. This is live. And as such, we would really encourage you to enter live questions into the Q&A session or the Q&A function on the left side of your screen, and we'll be able to address those as we move forward, both during the presentation in line, if I can answer those, or at the end when we have a formal Q&A, we'll address those then. I want to say any questions that we don't get a chance to address, if there's an excess of questions and we run out of time, I will personally guarantee I will respond to those when the session is over in the event that we run out of time and getting to them live during the session today. Joining me today for this session are Lin Nease, HPE Fellow and IoT Chief Technologist at HPE. Hey, Lin. John Magee, Head of Marketing for Digital Solutions at Hitachi Vantara. Hey, John. And Ishu Verma, Edge Computing Technical Evangelist here at Red Hat. Hey, Ishu. What we're going to do is we're going to go through some slides real quick. So the agenda for this session is very similar to the other sessions you may have attended during HCTS, like our Security or Data & Analytics tracks, short presentation, followed by Q&A with these industry experts. And then we're also going to have that time at the end for live Q&A with the questions you submit as well. So with the following handful of slides, what I want to do is walk through a few things. Since this is the first edge presentation of the day, I'm going to spend a little bit of time, framing the rest of the sessions for the rest of the day and mentioning those. And we're going to talk about what actually constitutes the edge. And so that means different things to different people, so we want to spend a little bit of time in defining that. The next piece is why people are adopting the edge. What we're seeing for enterprises in edge adoption and what workloads they're putting there and why. And then we'll get into the 5 key approaches, that is the title of the presentation, the 5 ways to Capitalize on the Edge Opportunity. We'll work through 5 key approaches in order to do that, along with the panelists and also with the help of yourselves during the Q&A section. So throughout the presentation, and I took some notes because I -- we have a very short period of time, so I want to make sure I don't miss anything important. One thing I'm going to do is I'm going to compare traditional IT workloads with the opportunities emerging as a key part of the digital transformation of industries and verticals. Sometimes that's called Industry 4.0, industrial automation, fleet telematics, grid modernization or the Internet of Things. While there are ample edge opportunities in traditional IT, if you can call it that, the industry that I've been in for 30-plus years, what we're seeing is a far more pronounced growth in industrial transformation areas, such as manufacturing, energy and utilities, health care, transportation, other areas that Eric mentioned and Simon mentioned in their introductory remarks. So we're going to work through all of those. And given the breadth of the industries that are served and the opportunities and the disparity or the differences in workloads in all of these sectors, it's probably no surprise that edge then means different things to different people in those separate context. That may be something as simple as the device edge, and that's far from simple, but let's just start on the left side of this slide. The device edge means things like drones or robots or cameras or maybe even smartwatches, things with embedded compute that may be doing their own analytics on board. Not all devices have that embedded compute capability and are capable of doing that analytics, which then lends itself to the next piece, which is enterprise edge. Enterprise edge is dedicated compute for edge workloads that might be on servers, in data centers, might be on robust gateways, micro data centers, those could sit in data center settings, nice temperature controlled, well lit, well ventilated. That might be on a rail out on a factory floor or in a closet or a cage or something else. So there's a whole number of different venues for the enterprise edge that, really, to serve, if you would, the brownfield environment. So all of the dedicated endpoints that are out there, maybe factory equipment or medical equipment, that isn't necessarily directly network connected or doesn't necessarily have the necessary compute or storage capability on board. The enterprise edge is the topic of a session a little bit later. My colleague, Christian Perry, Senior Research Analyst in our infrastructure team and the person voted most likely to accidentally receive e-mails, steps in for me on internal mail lists, will be moderating a panel at 2:00 Eastern Time to dive into more detail on the enterprise edge I would encourage you to attend. But there are industries and sectors that don't necessarily have the real estate on-premise or maybe they are assets in motion or maybe they're out in distant environments like offshore oil rigs or transmission towers on the electric grid, and that's when you see more adoption of network edge. So that's infrastructure that's going to be contained within network operators, carriers, and that's where you begin to see things like the forthcoming 5G, fifth-generation network build-out occur, as well as multi-access edge computing, which is compute, again, pushed further out towards the point of data origination, but in this case, within the network operator network, might be at the radio access network level or just slightly removed from that, again the purpose being trying to get the compute as close as possible to the analytics, to the workloads. The next step to the right of that on this slide is the Internet edge, and this is when things start to get super interesting. Oh, sorry, I should back up. The network edge is the topic of a really fascinating panel coming up a little bit later today from my colleague, Brian Partridge, the research director of our applied infrastructure and DevOps channel, and he has a presentation and some panelists that he's going to be going through during that session at -- in 90 minutes or so. So I would encourage you to attend that as well. Okay. So the Internet edge, not to get over my over-caffeinated tongue, it includes interconnection, multi-tenant data centers, content delivery networks, among other locations. So this edge serves a different set of workloads than the device edge or the enterprise edge or the network edge necessarily and is the topic of its own session led by Craig Matsumoto that Eric mentioned, our Senior Research Analyst on our Datacenter Networking team, and his session immediately follows this session and the short break that we have. And I would encourage you to attend that as well. That's when you begin to see things like CDNs come into play. So I would be completely remiss if I didn't state that the hyperscale cloud providers have been very effective at extending the benefits of cloud computing out to the various edges, and that really illustrates the larger takeaway of this slide, which is that the edge or compute, in general, for that matter, does not currently and will not in the future happen at a single location but rather across a continuum of locations from the machines themselves all the way out to the cloud. What this continuum provides or enables is infrastructure for a new set of applications to emerge, and there have been a lot of talk about augmented and virtual reality over many years. And we're beginning to see far greater adoption of that now than we have in the preceding 2 decades when it seemed like the carrot was just slightly ahead, and next year was going to be the killer year for augmented and virtual reality. We're beginning to see the adoption curves on both of those technologies grow very steeply in enterprise context, not necessarily consumer. We're beginning to see a lot more industrial automation, a lot more robotics and a lot more technology to assist. And maybe John can talk about this a little bit during his section, given Hitachi Vantara's presence in manufacturing, but really driven by COVID in the pandemic and distancing of workers on manufacturing lines in the -- and cobots, other assistive technologies in manufacturing. We're seeing a lot more cloud-based gaming for those of you that have maybe weathered out this pandemic by improving your gaming skills. We're seeing a lot more health care adoption of edge computing and being able to do things like telemedicine and remote monitoring of chronic conditions to avoid having to go into primary care facilities, so a lot of these are applications. Autonomous vehicles is sort of a perennial favorite when we're talking about edge-use applications and edge compute. There are a number of these applications that have existed in the past, but all of the necessary factors didn't line up as far as technology maturity, infrastructure, cost, bandwidth availability that are really lining up now to drive, a, these workloads; and, b, edge infrastructure. And those are being provided by, as I mentioned, a number of the companies represented by the panelists today: public cloud providers, private cloud providers, network operators, multi-tenant data center operators, colo providers, CDNs. And all of that infrastructure build-out that were just at the very beginning of -- if you think about digital transformation, we all have a tendency to talk about digital transformation in maybe a past tense or a little bit of a present tense, but really the majority of it is in the future tense. We're just beginning -- we're just at the very tip of the iceberg of global digital transformation. And as that build-out occurs, that's going to drive a ton of demand for -- well, for starters, it's going to be very large scale, so orchestration management at scale, a lot more dedicated silicon and specialized silicon for a lot of these different workloads, servers, storage, different networking technologies. You may have heard of like low-power wide-area networking or forthcoming 5G and then all of the associated infrastructure that makes that work, like racks and cabling and cooling and DCM. It's going to be a great explosion. It's going to be a great time to be in the industry, and I say that after having been in the industry for well over 3 decades. I've never seen anything at this scale and with this much potential to really upend how society works. Okay. So I mentioned a moment ago that I was going to talk about the differences between traditional, if you would, IT workloads and emerging digital transformation workloads. If you haven't attended any of the other sessions, let me just state real briefly. 451 Research, we have dozens of enterprise IT surveys that we perform multiple times per year named the Voice of the Enterprise across multiple technology spaces: storage, networking, security, Internet of Things, data and analytics. And we're fortunate in that we're able to talk to a diverse set of stakeholders across multiple job functions and industries. One of the by-products of that is we -- it begins to see or we begin to see how IT workloads differ in how they're being handled from new digital transformation, if you would, IoT workloads. And so I just want -- I have a couple of slides real briefly to walk through that. In looking today at where workloads are placed, the majority of IT responders, talking about traditional, if you would, IT workloads are delivering those on-prem, 55% of workloads on-premises but the remainder distributed across software as a service, infrastructure as a service, hosted private cloud, third-party colo facilities for good reasons, right? These are existing legacy workloads, traditional workloads, and they're in the midst, as you know, of being refactored and maybe shifted to the cloud in some cases, maybe kept locally for other reasons. We'll talk about those in just a moment. And that differs slightly from digital transformation or IoT workloads, and what we see there is edge computing, utilizing, depending on where you put that dotted line, 61% to 69% of cases where the data is being analyzed in an edge-ish or edge-centric type of environment. And it's important to mention that the emphasis here is on initial, initially stored and analyzed, because, more often than not, what we see is -- or are rather that these data sets might be initially analyzed. You might be taking in a large volume of time series data such as temperature data or vibration data or video, high-bandwidth video data, and you want to do the initial analytics. For economic reasons, you want to do that out of the edge, but you might summarize that and send that on then to further applications for secondary analysis or might be incorporated into other tools like CRM or ERP. So then, if you would, the new workloads that are being brought into these 2 -- across all industries, we're seeing far more adoption of edge in those cases. And it's worth mentioning that in both of these cases, the traditional IT workloads and the emerging DT workloads, survey respondents to both of these sets of surveys plan to expand their adoption of edge computing venues for the new workloads in the coming 3 years, both IT and new workloads in the coming 3 years that edge is going to actually grow. So why would you go through the trouble? I can just send everything to the cloud. It's so much easier. So the 7 primary reasons that we see people adopting edge computing, this is the enterprise respondents to our Voice of the Enterprise surveys, what they come back with are 7 key reasons. Let me go through these briefly, and then I'm going to walk you through survey results of where these land in and priority. So these are not in priority on this slide. So those are the need for low latency or ultra-low latency. There are industries that are real time, maybe semiconductor manufacturing or electricity substations, where you need to be able to send messages with synchrophasors -- phase management unit, sorry, very briefly between transmission lines and substations, so you can balance the electric grid. These are things that have very low latency. We're talking sub-50 millisecond, sometimes sub-20 or sub-10 millisecond application requirements. And they can't absorb the latency of a wide-area network connection linked to a distant data center and back before they can make a decision on how to rebalance a grid or to stop an assembly line without major quality impact or production impact. So low latency is sort of trotted out quite often as the primary reason for edge computing, but it turns out, it's not. It just turns out to be a benefit. A second one, one we see when we do economic analysis of workloads, is the cost and availability of bandwidth, so 2 and 3. A number of these types of applications, especially in transportation, energy, oil and gas, you might not have ready access to ample broadband for these applications. Or if you do, it might either be too slow or too expensive for -- to make sense for the workload, for the application that you're using. So cost of bandwidth availability and bandwidth turned out to be key factors in workload placement. And then the fourth one is something we're going to get back to in just a moment, which is the security. And you can argue this is either actual or perceived security of the workload in question. So I might have proprietary data. I might have very sensitive data that maybe even for regulatory reasons or legal reasons, I have to monitor or have to maintain the security of that data, might be video, raw video feeds. There might be privacy regulations or something like that. Or it might be trade secret data, production volumes that I really, really want to keep on-premise. That ends up being a key driver, and it may seem unnecessary. For those of you that are well steeped in the cloud, you know the cloud can be very, very secure as well. But there is a strong perception amongst enterprise respondents to our surveys that there is a security benefit to doing things at the edge. Another one of those that's closely related is data sovereignty. As I mentioned, there might be regulatory drivers saying this data must stay in this country or this county or this city, this province. And then there's always concern about what we used to affectionately referred to 20 years ago as backhoe fade, otherwise known as somebody severed my wide-area network connection. And so I need these applications, these workloads to continue to operate in a non-interrupted manner -- uninterrupted manner in the event of WAN failure. So I can't be dependent on that WAN link for my day-to-day operations. It needs to be resilient. That used to be called delay or even disruption-tolerant networking. And then something that's often overlooked is if you look at the brownfield, the installed base of existing equipment, much of that equipment has no network connectivity, has no on-board compute, so it's very important to have a bridge to these legacy systems. And oftentimes, edge computing devices can have interfaces to bridge those legacy proprietary maybe or legacy interfaces to be able to extract the data from pieces of manufacturing equipment or vehicles or something like that. So as I mentioned, and I'm going to go through this real quick because I want to make sure we leave ample time for the panel. What we see on -- from survey respondents on why they're deploying the workloads, where they're deploying them, the key driver for that is security. 60% of the time, that's the reason that enterprises say that's the reason why they're placing that workload where they placed it. So it's the driver for that where conversation and why conversation, followed very closely by cost, networking connections. Those are obviously unrelated, and then not surprising, availability of staff expertise. There's still quite a bit of a skills deficit across the board. Cloud, as you know, cloud skills are still in short supply or very expensive, again tightly correlated. But that's equally true in edge computing. It's equally true in wide-area networking. I may not have the skill set on staff to be able to deploy the workloads where I need to place them, and they may have always been on-premise. And therefore, I like it that way. And if I'm the one writing the checks, that's where they're going to be, and my team is already trained on that. Infrastructure resiliency, we talked about in the event of WAN failure, and then latency considerations, which, as I said, is often touted as the #1 reason for edge computing, but it turns out to be down the list as far as real-world selection of execution venues. Okay. So we're going to transition to the panel here. Just to frame that conversation, I want to do a real high-level overview of those 5 ways to capitalize on the edge opportunity, as I mentioned. And just real briefly, what those are is in delivering outcomes and not products. This is very much market transition, digital transformation, in general, a market transition that favors people who provide solutions to problems. They provide outcomes. And this is in contrast to traditional IT cells, which is me, as an enterprise, I want to buy best-of-breed components and then assemble my own gourmet meal from all of these great ingredients. Digital transformation is different, and we see this very, very clearly over now half a decade of survey results. And they want -- or what they're preferring to pay for, in most cases, is an outcome. I want better uptime. I want better quality. I want to know where my fleet is at all times, and that's what I'm writing a check for. And how you go about doing that, Mr. vendor, is on you. I'm writing you a check so that I don't have to do that integration exercise. And this is the reason that a lot of systems integrators and consultancies are thriving in this digital transformation environment. The second thing is designing for scale. Edge computing, as you would have imagine, grows very, very, very, quickly. I'm sure all of the panelists today will have great examples of deployments with thousands of endpoints. That's not something, speaking as a recovering network guy, that I want to config one at a time or deploy one at a time. I want to be able to design for scale, manage, orchestrate at scale. And as I mentioned a moment ago, there are skill gaps. Not everybody in the workforce, the IT workforce or on the operational side today is prepared for all of these things to come. The technologies are still being invented. The tools are still being created. And so not surprisingly, a lot of the people that have had those jobs are not necessarily ready and equipped to deal with the challenges that are to come. So there are skill gaps. There are existing channel partners on the light side of that, people that have been out there, working in these operational environments, working in these IT environments, hybrid environments that have pieces of that puzzle, that can help you, that can help enterprises deploy these technologies because they do have a lot of that domain expertise, a bit on the IT side or the OT side. And really, it's the combination of those together that creates that digital transformation. And then also, if you're a vendor in this space, something I can't amplify enough, is the need to build out a robust ecosystem of partners. Having that ecosystem will allow you to partner for adjacencies. And what I mean by that is there's such a diversity of use cases that you can't do it all. There's no one company that can do it all, and there are always going to be people that do just this piece or just this 5% or 10% or maybe 45% of that total outcome that we talked about that really need to be at the table with you in order to deliver the total meal, the total solution and assist in the transformation of clients. So we're going to drill in each of those separately with the panelists today. Maybe if I can ask our panelists to reintroduce themselves. Before I do that, the slide that I'm so excited -- I'm so excited to get to the panelists, I forgot to mention a very important point, which is if we look at our forecast and we look at the data coming back from enterprises, as far as the surveys and future workloads, what that informs us is that the edge opportunity is going to equal that of the current cloud opportunity, and it's going to track with it very, very strongly. There are scenarios where it actually goes faster, but we're just going to be conservative and say it's going to track with the cloud opportunity as these -- as digital transformation hits more mainstream industries, things like manufacturing at scale, transportation, energy. So now with that, I'm going to transition to our panelists real quick. Maybe we can do a lightning round of first introductions, and then we've got some questions along those 5 trend lines. This is another point where I would encourage you to avail yourself of the Q&A utility in this webinar tool. Please have your questions in there, and we'll definitely attempt to get to those in the time remaining. We'll go through all of our questions as fast as we can. Lin, can I impose upon you first? Maybe introduce yourself and the company.

Sure. So I'm Lin Nease. I'm the Chief Technologist for IoT and what is HP Enterprise's Pointnext services, advisory and professional services business. So I spent a lot of my time, both doing strategic consulting for key customers along these areas as well as driving some of our internal strategies in HPE.

Awesome. Thanks, Lin, and thanks for joining us today. John Magee?

Yes. Thanks, Christian. Hi. This is John Magee, and I lead go to market for our digital solutions business unit at Hitachi Vantara, and that includes our Lumada IoT platform and portfolio. Pleasure to be here.

Awesome. Thanks for coming, John. And Ishu Verma from Red Hat. Ishu?

Thanks, Christian. Ishu Verma. I am the Tech Evangelist, working on emerging technologies, so I get to work on exciting things like edge and IoT. And I've been doing IoT for a number of years, but, since last year, edge is becoming more important for our customers and excited to see how the change is reflecting in the marketplace with our customers.

Excellent. Thanks, Ishu.

So let's kick it off. Let's walk down maybe that list of 5 ways to capitalize. Lin, if I can start with you. Pointnext, HPE. How are technology providers, I guess, adapting to provide those outcomes versus piece, parts and products? I mean I think our industry, traditionally, has been structured to deliver a box or a component of a total outcome. How is that pivoting to providing more of an outcome in your opinion? And maybe any examples you have from HPE.

Absolutely. So I can give you a few examples. But to start with, the edge, in many cases, represents a place where IT skill sets are not in strong supply. And when you take into account the transformation use cases that the edge is clearly bringing about, video pattern recognition, new types of analytics that are running for condition-based maintenance and the like, what you find is a huge skills mismatch and a lot of complexity. There are a lot of different types of vendors and a lot of different types of technologies and no customer of ours, and I'll describe some in a minute here, have -- are equipped, quite honestly, to do a best-of-breed type purchasing cycle here. And what happens with this complexity is you kick off a project. Let's say I want a roll out a new use case. And in this case, we had a customer, 2,300 stores, a very large home improvement retailer. I wanted to add a few use cases for running retail analytics, in addition to hosting their traditional in-store, point-of-sale inventory, crew management, we'll call it, as well as enabling some location capabilities in their network, right? So the cone of potential outcomes, I'll call it, like the -- if you just run the analysis, not knowing how this project is going to go and not having solved all the details of a very complex edge project, the cone of potential outcomes was too great. And for this customer, it could easily cross in a negative ROI very easily for a very large chunk of this cone of potential outcomes for the project. So with our expertise, right, we actually have built our entire company strategy at this point around something we call GreenLake. GreenLake is our as-a-service, outcome-based basically service portfolio. Everything in our company now is moving towards as a service and annual recurring revenue. And in this case, what an outcome means -- don't go crazy on the derivative stuff. I'll tell you about that in a second. But what it means here is it's a futures contract to the customer. It's like, look, I can't tell you how many people, how much time, what these different project steps, there's too many things beyond my skills. I need to lock in a price, "lock in a price." So what we've done is created, in essence, a futures contract approach to how our GreenLake services work. We have the tool, and we have enough, I'll call it, whole value chain tool. Everything from we have a bank, we have recycling capabilities, we can buy back assets, we can modulate what functionality gets rolled out when, and we have great project management, right? So because of our ability to modulate on our side, we can absorb the risk. We know enough about the customer's problem where we can offer them a price based on square foot, based on number of employees, based on number of small, medium and large t-shirt, sites, for example. Is this a 2-server site? Is this a 5-server site? And then because of our ability to modulate and take care of a lot of the tasking that our customers -- it's a checkerboard, too. In some cases, they have skills and not in other areas, but their ability to lock in a price based on something they know that can generate a positive ROI for is what this outcome is about. And some people have gone crazy with, I'll call it, derivatives pricing, where I'll take a percentage of your savings and the like. To me, that's very, very difficult to pull off because procurement in our customers don't know how to really price for risk. And then what happens is they RFx you to death so that it's very difficult to protect the IP that allows you to offer that price. And then the customer can turn around. And not to say customers are evil for doing this, they're not. They can look at how you've decided to roll out the project, learn from that and then turn around and bid it out the way they always have. So the far you get into this crazy zone of, I'll call it, derivatives pricing, the more difficult that is to operate as a business. But definitely, the ability to modulate and run a project is of great value to our customers. And at this particular customer, 2,300 stores, they had at least 2 servers in every store, and they needed a new wireless network in many of them. We did a combination of asset buybacks, recycling, rolling out in a delayed form for some of these stores. And in 9 months, we have completely revamped the networks and the server infrastructure in all of those stores. And what we did, the outcome here was there is a monthly price that is being paid today for the combination of leases on the equipment that they expected to get replaced. We stayed within the balance of that monthly cost. So basically, the perpetuity rolled right on through it with the same OpEx. That's new network, new infrastructure, new use cases.

Okay. And so I'm going to circle back to this in just a moment, but thank you. I appreciate the example, the real-world examples as well. I think that's important for everybody who's attending today to understand, too. As I say, look, they hit the berries from that bush, and they didn't die. So it's okay for us to eat the berries, I think, all these examples.

Exactly. And honestly, these conversations get into the sexy stuff, the new use cases and the like, but some of it is nuts and bolts. It's fundamental absorbing project risk.

Amen, amen. Yes. We have evolved our surveys in the last 5 years to go from, what I call, duckies and ponies, the really early stage. Are you doing these things? Is it important to you? Now we're talking about how do you capture ROI? How do you measure? What is your exit criteria from trial into production, scalability, things like that? Because it is a maturing market, and I think now we have the opportunity, as an industry, to point at successful deployments and say, "Here's what worked, and here's what didn't," and save others the same headaches and loss and budget headcount, so forth. Ishu, can I ask you a question? The -- we talked about scalability and manageability. And obviously, this is something that just blows out, I'm not going to say exponentially but darn near when we talk about edge computing because we -- as Lin said, thousands and thousands of locations. This isn't something I want to manually command line into each box. How are we going from proof of concepts in small trials to scale, large-scale production deployments? What are the things that are necessary in order for that to actually work in contrast to what we've been doing, frankly, in the industry for many, many years? How is this different?

Yes. So we are seeing, Christian, customers' requirements have evolved in the last few years as they move from POC to mass deployment. And scale in this term is obviously the number of locations that edge brings, tens of thousands of locations but also the volume of data being generated on the edge. So I think that's also part of scale. So one example I can probably think here to reference to the number of locations is for oil and gas industry. So Exxon is looking to modernize their scalar system a few years back, and they decided to go with bringing IoT gateways to connect with their brownfield infrastructure. Very typical approach, right? You already have existing infrastructure you want to connect and bring data from those systems. So they chose an off-the-shelf hardware, consumer-grade operating system and had really quick POC out of the gate that had continuous applications running on it. They were doing instant upgrades, and they were rolling them out slowly in the field. And they got, I think, all 700 gateways. And 1 day, as they updated the system, it breaks the gateway. So Exxon had to actually do truck rule to replace these 700-plus gateways. So they [indiscernible] actually.

Because they are all I care for.

Absolutely. And they -- I mean they were very clearly on what their requirements were. So they were looking at intelligent rollbacks to avoid the situation they had there. If the system is updated and the update is not successful, it should be able to roll back to the known good state. So the second requirement was transactional cost update. So they were looking to see the -- these gateways are deployed in locations that have poor network connectivity or really low-bandwidth network connectivity. So you cannot expect 100-megabit packages being downloaded on these wires and updated. So they wanted thin layer based packaging update being provisioned. And then the third requirement, key requirement was one-touch provisioning. I think you talked about, Christian, how these sites are now going to have IT skill sets. So at these sites, you will have a low IT scale or somebody who is just going to plug in the box and press a button. That's all they can do. So these should be managed centralized, automated way, and it has to be one-touch provisioning. So essentially, for them, the key learnings from -- as you move from POC stage, you have to think about the large-scale implications and not to POC and then worry about actual scale that you try to deploy. On the other side of the data side, where we are talking about scale at data side, there are some things, like BMW, who are doing autonomous vehicle development. and these generate huge amounts of data. I mean you have all these sensors and LiDAR, generating almost 8 terabytes of data an hour, so several systems. They are talking about terabytes of data a day, right? So on top of that, you need to be able to update these applications and run analytics models on the factory to respond to this data. So for them, they are looking for environment they can run these tests in real time, scaling in autonomously, right? So it's rough, heavily develop, furthering their own containerized environment. So they actually chose to extend their centralized infrastructure capabilities using communities to run at the edge, and then they can do the -- all the great things that it brings, automated CICD pipeline, doing -- updating these algorithms and modes at the edge in real time. And we are seeing actually other industrial manufacturing companies bringing enterprise communities to the level 3, right, so this contract volume model. So they can actually process these at the edge, at the factory floor, at least, using the typical tools and methodologies they used for their centers.

Okay. Thank you. That's a great -- a couple of great examples. Thank you. Mr. Magee, I guess probably pretty timely, given the relatively recent joining of Hitachi Consulting into the Vantara team. Skill gaps and consultative sale processes, how has it changed, I guess, the way -- how is this different in this -- I guess the sales action or the go-to-market action that you've seen, especially given your previous experience with other companies? How is edge and these digital transformation processes, how are those different in how they're sold and deployed and created, cocreated versus in previous times I guess? I've definitely seen a change myself. It used to be you leave it at the dock, and you need -- sign off, it got delivered. Now it's sign on accept, not sign on delivery.

Yes, yes. I think that's right, Christian. The go to market and the way you engage with customers is different, partly because of the complexity that we've been talking about, this continuum of edge deployment. It's complex, the topology that you might end up in with the solution architecture. But I think there's a bigger issue, which is, really, we talked about outcome. So you mentioned this is digital transformation, which means people are going to do things in a different way than they've done them before. And so just defining those outcomes can often be an opportunity and a challenge in the sense that you're coming into these environments where people have different opinions about what they're trying to accomplish. And usually, there's a pretty good high-level focus. And of course, everybody wants to get to that solution architecture and deployment plan, the POC and everything, but there's also a lot of opportunity that we found to really work with customers on refining that, sometimes competing different agendas in terms of what those outcomes should be and how to get there. So we've gotten pretty formalized now around a consultative sales process involving, we call it, co-creation, where we've got a methodology and involves design thinking. But the real key is bringing together multiple stakeholders. And I think this is not new. We've gone through all these iterations of -- generations of technology from ERP to the Internet to e-commerce, and those were similar kinds of changes. But I think the big dynamic here in this space with edge and IoT and the industrial sectors is the OT and IT dynamic, we'll call it, and that varies. And it brings in a lot of complexity sometimes where you've got to get both sides working to get consensus around what the objective should be. And also, it helps to understand it, and it's always different. Sometimes large manufacturers, IT is taking the lead. Sometimes it's OT in different industries. And so we really see this as a tool to kind of bring the stakeholders together around the definition of what the solution should accomplish for the business, and then you can get into all the how do you phase that and stage it. We've done this at Disney. If you think about Disney, it's not a traditional industrial company but lots of different stakeholders, and we've been working with their Parks group. They were a long-time customer of Hitachi Vantara's for infrastructure, storage, compute and so on. But on the IoT side, they're really engaging with their parks. And they had a lot of ideas and their teams had a lot of ideas, and these teams were very diverse. There's reliability engineers just like you and me in a manufacturing environment. There's the kind of big-picture folks that are doing the animatronics and everything and so a lot of creativity. But bringing all those different stakeholders together to get to a consensus of where to go next is something that I think is an opportunity and a challenge that needs to be addressed. And again that IT, OT dynamic, we see this in a lot of engagements, navigating that and figuring out who the stakeholders are as key.

Well, hopefully, we can all get to a good theme park again sometime in the near future and not just look at them through a web browser. Maybe a follow-on question to that and to yourself and to the other panelists. Have you seen that co-creation activity? The design thinking deployment process or sales process, how does that bubble up skills gaps between maybe the IT and the OT constituencies or anybody else who is sitting at the grownup table when these things are being designed and decided upon? What have you seen as far as just glaring -- there's a whole -- there's this half of the team, this half of the team, but there's all this white space between them.

Yes. I mean there are still skill gaps today. I think a lot of our technology base, like you touched on, just having the expertise in some of these things or not having the staff to work on some of these things. So customers need help with that. But again, from a co-creation standpoint, I think there's also a skill gap in thinking about what's the art of the possible and really thinking bigger. Frankly, a lot of times, customers are kind of starting tactically. And if you can envision a different way of doing things, that's a big opportunity. The other skill gap that we see is change management after it's installed, right? Traditional markets like ERP, it's very common that you're going to do a big ERP rollout. There's also a post-project change management, where you're teaching all the people involved how to take advantage of these new systems. That's not really formalized very well yet in the industrial IoT space, and I think it's something that's going to have to happen. It's something we're working on. It's like you've got now analytics to solve problems in new ways than you had before. How are you going to get these new -- these people adopting these technologies, working in different ways, collaborating differently? So that whole area is one that I think is another opportunity for skill gap closure.

We see huge skill gaps in the OT space for things that in the IT world are pretty straightforward like security operations, cybersecurity. And as -- I mean everybody here knows this. This is a huge issue now in the OT world is systems are connected that used to be air gapped, right? WannaCry came along a couple of years ago and taught everybody that you're going to have to learn how to patch some of these antiquated OT systems or rethink how they are connected on networks, and skills just aren't in the OT organization.

Ishu, are you seeing similar things on your end? I mean do you run into the OT folks and the IT folks in the same sales process? And...

We do. But I think, increasingly, what we are seeing, especially with the context of edge, that IT is, I think, more at the decision table than they were before with the IoT days, right? They are -- I think what you are still seeing is, this is my layer, and you do not come in and tell me what system. I will manage this, right? I'll make decisions, but I think the -- we are seeing the idea that we're extending from core to the edge on a horizontal, consistent approach. And then IT -- OT workloads can run on those, but I think IT is making some of the decisions, what those and infrastructure environment is going to look like. And security, I think, Lin, you talked about. I think they understand it much better, so they are saying we'll take in charge of that. You are just one of the [indiscernible] on this running. And -- yes.

Okay. Well, great. We do have additional questions, but I've just been told that we are out of time because, evidently, I got too excited on my introductory slides. I'd like to thank you, the panelists, again, thank the attendees for attending. I will be in the virtual lobby, I guess, or there's a facility for that. So if you want to do some chatting after this session, I would encourage you to do that. But thank you again to the panelists for sharing your wisdom and your real-world examples and watching my dogs snore in the background. So thank you, everybody. Enjoy the rest of your HCTS experience. I would encourage you to attend Christian, Craig and Brian's presentations during the rest of today around the edge and explore the other assets of the edge. So thank you again.

Thank you, Christian.

Thank you.

Thank you.

All right. Have a great day, everyone.