Kelly Lynch: Hello, and welcome back to the Next Horizon. Per usual, I’m Kelly Lynch and this week I’m throwing it back to the interview format. Hope that’s cool. We’ll talk all about predictions for 2021, how technology is going to continue to evolve, and what you can really expect to see from some of these technologies next year. Whether you’re driving somewhere, relaxing in a cozy chair with a big sweatshirt on or doing the dishes, which is my personal favorite podcast listening activity, I hope you’ll enjoy.
Kelly Lynch: Hello everyone and welcome back to the Next Horizon. I am very lucky this week to be joined by John Roese, the Global Chief Technology Officer of Dell Technologies. This is actually his second time with us this season. John, thanks so much for coming back and spending some time.
John Roese: Glad to be here.
Kelly Lynch: I guess, to start, I’m not saying anything new that hasn’t been repeated on probably every podcast or every TV show ever, but we are and have been living in a strange time. Our work and home lives have become much more integrated, a line that’s admittedly a little bit blurred. But for the most part, we’ve adapted and this adaptation comes a lot from human resiliency, but also a lot from technology. We’re experiencing this human transformation, if you will. I’m curious, to kick things off, what are some of these key technology inflections that will continue to enable this mixed reality world that we now live in?
John Roese: Yeah. When we talk about mixed reality, which sounds like a science fiction term, all we’re doing is just saying, “Look, the human experience is now merged and blended with the digital experience for us to be able to get work done.” If you want to do work during COVID, the odds are pretty good you’re going to be using technology to do it because you can’t go to your office anymore. Your place of business might not physically be open, but the actual business that you’re doing could in fact be achieved by moving through a video conference or engaging over maybe a virtual reality environment. There’s lots of different choices, but the bottom line is that that mixed reality experience is very dependent on technology and the technology is still evolving. And right now, we’re in a pretty rich period of evolving technologies.
John Roese: We have the entire hybrid and multi-cloud environment materializing where we’re realizing a single cloud is not enough. We have to have multiple clouds working together as a system to deliver the platform to do all of this stuff. At the same time, we’re realizing that the client device, the worker itself is not independent. In fact, your device at home is a portal into a much bigger digital ecosystem that involves edges and core data centers and public clouds and all of the services and that concept of hybrid IT and the hybrid user, that not everything is done on your device, but your device is incredibly important to access all of those other things is starting to materialize as a technology trend leading us to things like collaborative compute and other paradigms that are going to emerge over the next several years.
John Roese: We’re seeing the delivery of IT services in real time out at edges that are forming. We’re seeing mobile connectivity be transformed into the 5G era, which is just beginning, but it has a lot of promise to allow us to maybe have this conversation anywhere, not just in a house with good broadband, which would be nice. And then beyond that, we’re dealing with new classes of users with artificial intelligence and machine learning becoming a dominant pattern of application delivery and data processing. And then on the data side, we’re seeing the entire data world shift from data at rest being analyzed to data in motion over pipelines, having a real time impact on whatever process that data is involved with, whether it’s a process line in a factory or an autonomous vehicle, or any number of different outcomes.
John Roese: And then lastly, we’re having to reinvent the security ecosystem because, guess what, as we create this digital value in this digital environment, it creates an enormous target with a lot of risks. And so, we have to continuously stay ahead of the security curve. There’s a lot going on right now, but it’s all in service of one thing and that is to keep humanity productive in a period of time where we are maybe isolated from being able to do things physically.
Kelly Lynch: Yeah. I mean, absolutely. I certainly myself have struggled a little bit with productivity, but have been enabled by all the pieces of technology that I have. But you’ve touched on a lot of different things in that opener, which is amazing and I want to try to break down some of those. To start, Dell talks a lot about hybrid cloud and we have. We’ve spoken about multi-cloud for many, many years. Now we’re seeing this hybrid cloud for the hybrid workforce. What does that mean? Could you elaborate a little bit about that for me and what it means also for IT?
John Roese: Yeah, yeah. I don’t want to say, “We told you so,” but for eight or nine years back to EMC, we have been unambiguous about the fact that your IT environment is complex. It needs to have the capabilities that exceed any particular one place in infrastructure and as the cloud reformed early on, there were many people that believe that cloud would be a single, monolithic, data center somewhere else that everybody would just use and there’d be no IT anywhere other than this one blob of technology somewhere. And that arguably makes no sense, but five years ago, even three years ago, maybe even last year, people really believed that was possible. Today, no one believes that. Even the public clouds, as of yesterday, they’re standing up and saying, “Nope. Hybrid’s the answer,” and what hybrid really just acknowledges is, “Look, you are going to need IT where you need it.”
John Roese: There are some applications and workloads that make a tremendous amount of sense to do in a very centralized, public, shared environment, i.e, a public cloud. There are other workloads that absolutely are so sensitive or important or performant that they need to be on a dedicated infrastructure that you can optimize and own and protect. There are other workloads that need to live very close to where the real-time action is, i.e, on an edge. There are even other workloads that absolutely belong on your device and they should live there and that device could be a laptop, but it could be an autonomous driving car. And that car needs certain things to be always present, even if it’s fully disconnected. And so, the idea of hybrid is just a topology. It says, “We’re going to have a distributed system. We’re going to have different places where we can use and deploy IT capability and the key is to make it work like a system.”
John Roese: The real answer is, “Look, hybrid IT is just simply this concept, whether it’s applied to the cloud or the user, that these outcomes require more than one infrastructure. They require the collection of IT capabilities to work together as a system and in many cases, a single workload may in fact, touch many points within that hybrid continuum. Again, we have been saying this forever. We completely believe it. It’s nice to see the industry come around and acknowledge this. It’s actually quite a good thing for Dell because we are experts on on-prem infrastructure, edge infrastructure. We even build public cloud infrastructures, but the idea is we’re an expert on infrastructure and what hybrid says is, “Infrastructure is going to be in many places. It’s going to need to be consistent and it’s going to need to operate as a system.” All things that we do really well,
Kelly Lynch: I say, “Toot that horn.” But with everything that you just said about the hybridization of work and how everything’s going to be interconnected and coming together wherever you are, it seems like, you mentioned it, everything truly is moving to the edge. Could you help me understand if the edge is a new perimeter for the cloud? And if so, how does this relate to multi-cloud?
John Roese: Yeah. Well, first of all, not everything is moving to the edge. That’s one of the things that we’ve had to work through because first of all, we used to have edges. Long, long time ago, when dinosaurs still roamed the earth, we used to put email servers out in factories and under our desks and maybe we’d call that an edge. The IT was very distributed. That was really dumb. Those workloads don’t belong at the edge because they don’t really have a need to be there. And so, we centralized all of those. In the new edge that’s forming now, really, there are four things that justify putting something into this new edge. One, if it’s a real-time processing decision. If the time to make a decision and process the data is so fast that you can’t afford the latency of moving across the internet, then you need to do that task at the edge.
John Roese: The second is data pipelines are getting so big and so complex. The amount of data we’re generating is just enormous that if you try to move all that data across the internet and then process it, the internet might not be big enough to do that. The third is the IT/OT environment. This world IOT that materialized was really this idea of, “We’re going to have machines and physical devices that need to connect to the digital world.” Where do you do that? Well, the best place to do it is as close to that device as possible, to convert between crazy protocols like Zigbee and these legacy protocols like SCADA networks use and turn them into internet protocol and make them part of the IT system.
John Roese: And the fourth is that the edge is actually becoming a very powerful tool for security because if you push the security functions close to the edge, you reduce the surface area to attack. If you have a sensor that’s being protected by a piece of code that’s literally on the other end of the internet, you have a very large surface area to protect. If on the other hand, you can take that security function and not move it to the sensor, but move it to the thing in the factory that represents the digital edge, now you have a very narrow area of attack surface and you’re going to reduce the probability of a failure or a catastrophic event significantly.
John Roese: You got these four real reasons, why you do edge and generally speaking, if it’s not one of those four, you probably shouldn’t put it on an edge. The good news is those four are huge. The world is moving to real-time data. The world is creating more data. The world is connectorizing and censorizing everything and the world is very nervous about security. And so, as we put those four things into practice, we estimate that up to 70% of the world’s data is going to be created outside of a data center, public or private, which means it’s going to happen at the edge, and that is an enormous part of the IT ecosystem. We’re just beginning. That hasn’t happened yet at scale, but anywhere where we start to see edge deployments, where people are extending their IT topology to be able to do IT in the real world in real time, we’re seeing the data and the application demands explode.
John Roese: Now how this relates to the multicloud world is edge is not a separate entity from the rest of the multicloud. It is just another part of this hybrid cloud that we’re building. Today, we think about hybrid as public and private working together, but we have to start thinking about it as public working with private, working with edges and working with telecom clouds because they’re all part of this end to end multicloud topology. And the reality is if you don’t have a common control plane, if you don’t have logical connectivity with SDNs, if you don’t have a pane of glass to visualize and analyze and understand the end to end multicloud, including the edge, you will create a set of silos that actually won’t be able to do hybrid IT, that the workloads and data pipelines won’t actually be able to execute because they won’t be able to actually use all the services of the different cloud topologies as a system.
John Roese: It’s a very big deal. It’s likely to be the largest area of on-prem “out of the data center infrastructure” in the world. It’s going to be where most of the data’s created. It likely is going to be a very high value creation area. But most importantly, it is different than building a data center.
Kelly Lynch: Which brings us to something. I mean, you said “70%,” if I’m recalling correctly, “70% of the data at some point in the near future will be processed at an edge.” That plus the fact that we’ve moved to this new hybrid, working hybridization of work model and we’ll continue to do so. That means connectivity becomes that much more important. And for people who have listened to the podcast before, they know that I have not the greatest of internet. And so, sometimes I struggle with Zoom calls like this because there’s latency and lagging and all of that stuff. I struggle with some of those things from home. I’m excited about what’s to come with increased connectivity and I know that you’ve said 2021 is going to be the year that brings about all of the cool things that 5G will enable. Besides higher speeds and lower latencies, which I certainly will benefit from, I’m curious, what other cool stuff you predict we’ll see, especially for enterprises and then how we, as a company or how Dell as a company, is going to enable that?
John Roese: Yeah. Yeah. And first of all, I actually don’t have to predict it because it’s actually written in the standards and is very, very real, just not available [inaudible 00:12:06]. It’s going at least to predict the future when it’s already been written. That’s how standards sometimes work. The bottom line is, 5G is not just 4G on steroids. It is a much more sophisticated and advanced system and it has some significant differences. The first is that 4G primarily was about delivering mobile broadband, that internet over the air, and it did a good job of that, honestly. If we didn’t have LTE we would still be running very lightweight applications on our mobile devices and probably typing on number pads and it would’ve been a nightmare. The smartphone was enabled because we had mobile broadband and then that opened up this very rich, mobile experience for applications that you could run on your desktop. You could run on a laptop, but now you could run it on a mobile device, and the only thing that it did that enabled that was you had a pipe over the air. That’s a good thing. It was a very important step.
John Roese: Now, 5G doesn’t lose that. In fact, it does make it best or better. The first characteristic of 5G is enhanced mobile broadband, which basically means, wider channels, more bandwidth, lower latency. Yeah, the mobile broadband experience will get better. Sometimes you joke, “YouTube will run faster. Yay.” If that’s all it was, that would be good, but that’s not anywhere near all of what 5G is about. The biggest difference with 5G is that much of the other technology was built not to make YouTube run better, but to make enterprises run better, to make transportation networks work, to make factories run better, to make autonomous vehicles work. And there, we see a whole host of new technologies that with what’s known as release 16 or release 17.
John Roese: The newer standards in 5G are starting to materialize in 21. We’ll start to see the first variations of those. They include things like massive machine type communication and the ability for 5G to effectively handle dense sensorification of the world. The idea of a million sensors in a kilometer and being able to understand, connect to them and manage their power effectively so they can actually survive in that environment. You can’t do that on WiFi. You can’t do that on 4G. You can do that on 5G as we move into these later releases.
John Roese: The third characteristic is something called ultra reliable, low latency communication. This idea of, “Well, what if we want to actually give a special quality of service or experience to a workload?” Maybe you have a drone with real-time telemetry feeds coming in, or you have a control system that’s super high bandwidth, or maybe you have a, I don’t know, a holographic immersive experience that you’re using for real-time surgery. Those are big deals and right now that’s not best effort. That has to be special. And so, ultra reliable, low latency communication gives us the ability to do that. All of those things are new. We didn’t have them in 4G. We have them in 5G. It’s going to have a lot of impacts. I can’t tell you exactly which one’s going to be best, but this is one of those horizontal changes that if we get it right, we’re going to overcome one of the biggest problems we have with cloud IT and IT in general, and that is relatively weak, scattered, fragmented connectivity.
John Roese: You get that issue off the table and it unlocks tremendous value about how you do hybrid IT, how you enable the user, how you work from home, how you do digital healthcare, how you do digital transformation and how you do anything that a business is trying to achieve. It’s a big deal, but we’re on year number one and a half of a 10 year cycle. We have a lot of time to actually work through this and continuously improve it.
Kelly Lynch: Those cool things like remote surgery, which I mean, moving from talking to a healthcare provider on my computer camera and having those feeds be faster and there’s no lag, cool, that’s a baseline. But moving from that to remote surgery, that’s a huge leap and if that’s what we’re seeing in year one and a half, I can only imagine what we’re going to see in year nine and a half so hopefully… I mean, what could that even look like?
John Roese: Well, we know that… I joke. Here’s an analogy to think about in terms of the current experience that we’re having right now on this Zoom call. We are attempting to recreate person-to-person communication. That’s we’re trying to do. I mean, honestly. If you and I were standing in a room talking to each other, we’re trying to emulate that with the tools we have. Clearly, we have not been successful. We are good enough to be functional, but we have a lot of work to do because when human beings are standing face to face, what happens is not only do they have the two-dimensional imagery like we see here and the audio streams and the visual cues, but they also have depth sensing. They have emotional connectedness. They have the environment and the ambiance around them to help influence the behavior and the experience they have. They have context around them. They understand peripheral vision. We can see. We’re able to share joint experiences around us.
John Roese: And every one of those additive things that make the human to human interaction when it’s face-to-face work can be translated into more data. You simply need more data and more capacity in the digital experience to achieve that equivalent experience of two people actually talking to each other face-to-face. As we do 5G, as we build out these cloud environments, as we expand the capability of the IT infrastructure, what inevitably is going to happen is the digital experience is going to get a heck of a lot closer to two people having a conversation with each other, and then it’s going to exceed it. The closeness is going to happen with things like holographic presentation, with depth sensing, with immersive environments, which we’ve already seen to some degree with AR and VR, but we’re going to exceed it in interesting ways because now if we get to the point that that experience is almost exactly or equivalent to what we can get face-to-face, but it’s digital, we can insert other digital value in it.
John Roese: Imagine us having this conversation. It’s not just you and I, but it’s a number of AIs that are participating in the conversation, that as we’re talking, they’re taking notes. As we’re talking about information, they’re going out and finding graphical representations to show pictures of what I’m talking about in real time without me telling them to do it. They’re collaborating with each other to understand context. Maybe they’re translating everything we’re saying into other languages in real time or the other way. It’s going to be a fun journey but honestly, the best thing that’s happened to us is people being forced into the digital experience like Zoom and other ones because they’ve opened up their eyes to be able to participate in that platform and we’re going to make that platform better.
John Roese: But then we’re going to make it a platform where we can aggregate other digital intelligence and exceed what two human beings standing together could ever achieve and that is a very powerful vision. It will take us a long time to do that, but if you want to measure success it’s, “How far are we on that journey? Have we gotten to the point that it’s equivalent to two human beings?” And then more importantly, “Have we exceeded it,” and if we’ve done that, then we’re actually in the future.
Kelly Lynch: We’ve got edges hybridization of work, if I can say that word correctly, hybridization of work. We’re talking about ARs, VRs, mixed reality, 5G, increased connectivity, this connectivity network and fabric for all of these things to become a true, scalable, continuous reality. I’m curious what other areas of technological change we’ll need, or that you see from a compute perspective or from quantum perspective that we’ll see coming forth in 2021.
John Roese: Well, quantum’s a different discussion. We’ll talk about that in a second.
Kelly Lynch: We’ll talk about that later. Okay.
John Roese: But which is worth talking about because it’s got a lot of hype. There’s a reality there. The most important change that’s going to happen in 2021 that’s been going on for a while, but we’re going to start to see it in a more pronounced way, and we’re already starting to see it in the last several months, is the semiconductor ecosystem changing. We have largely operated in a world where we have the benefit of x86 standardized compute, and that will not go away. In fact, we’re still going to run most of our code on x86 processors. It gives us a nice, standardized layer to basically develop software and put it into production. And so, the semiconductor ecosystem is actually in a fairly aggressive innovation cycle that’s actually creating a heterogeneous compute environment that is not compute done one way. It’s compute done many ways in the infrastructure.
John Roese: The advantage of that is that each of those ways, especially in the accelerator space, can give us two or three orders of magnitude improvement in mips per watt, or performance for tasks, or whatever measure you use to judge the capacity or capability of that particular chip. That’s a very powerful tool because what it allows us to do is to run our IT infrastructures far more efficiently to get to task completion much faster. The cost of that though is when you move to a heterogeneous environment, it adds complexity. These things are not the same. The software needs to understand that or be hidden from it. The users need to actually deal with diversity, which is a really bad thing, in some cases, when you’re trying to control complexity. But the inevitable trend here is with that diversity, we will get compelling performance and outcome advantages.
John Roese: We’re going to have to overcome the complexity by better system architecture, by different relationships. And so, the semiconductor ecosystem is going to change and evolve and become fragmented and heterogeneous. The two ways to deal with it if you’re buying product, you need to buy from people who know how to put these systems together. Dell’s quite good at that. If you’re buying services, you need to understand that those services are likely going to be quite agile underneath them adopting these new technologies which should set an expectation in most customer’s minds that there is a more rapid innovation cycle, that performance, cost and functionality should actually move faster in the as-a-service market than it does in the standalone product market. And while that sounds obvious, that’s not entirely true when we look at the current as-a-service market with public clouds, where they offer a lot of new services, but the services are relatively static once they’re put into production in many cases, and we actually think they can move faster there.
John Roese: We think we can move faster as Dell, but the bottom line is the underlying inflection that’s causing these changes starts with this heterogeneous compute paradigm, this idea that we’re not going to have a single architecture at the semiconductor level. We’re going to have lots of them and we’re going to have to make sure not to exploit them the customer doesn’t have to go through the pain and effort of having to deal with that variability. They can still just consume a product. They can consume a cloud. They can consume an edge. They can consume a multicloud or they can consume a service and all of those abstractions are going to be critical to keep this complexity that’s actually helping us improve the performance and cost of infrastructure away from the customers who ultimately have to consume it.
Kelly Lynch: And we did mention quantum before so I want to pick your brain a little bit about this first. Can you help me understand a little bit more about it, when we expect it to “Be here,” and what real value it’ll actually bring?
John Roese: Yeah. Yeah. First of all, quantum computing is not a new concept. The physics behind using quantum characteristics, which are very real, superposition, the other effects that are absolutely true in physics. There’s no black art here. These are things that we are unlocking the mysteries of the universe in the physics level and we’re discovering that quantum mechanics and the technologies that exist around it can actually be very useful. We’ve already used quantum mechanics to build better materials, new lots of things in our life that we already touched. Battery technology is impacted by quantum, but quantum computing is this idea of, “Well, there are certain characteristics in the quantum environment, specifically the fact that when you excite a piece of matter to a certain level, it behaves in a way that’s unusual in the sense that it can exist in many states simultaneously. As opposed to in a binary world, it’s a one or a zero.”
John Roese: And that fundamental effect within the quantum environment when you apply it to compute, gives us an entirely different way to think about processing data. Now that is incredibly specialized and incredibly hard to do and that cubit, which is that unit of energy or matter that’s being excited, that can be in many of these states, has to be created. It has to be maintained. And unfortunately, it’s very hard to do that. And today, the biggest quantum computers in the world can produce maybe 70 cubits. And the reality, to do anything interesting, you might need thousands or millions of cubits and those 70 cubit systems cost tens or hundreds of millions of dollars to build and they have to sit there at absolute zero, and they’re incredibly complex to deal with. All of that should tell us, “Well, maybe this isn’t going to matter,” but the reality is they’re just technical challenges.
John Roese: We have to do three things to make quantum computing more viable. The first is we have to define and architect a viable quantum computing system. We have to figure out how to make one of these things and we have all different experiments going on. None of them look like a fully scalable quantum computer, but we’ll get there. We will figure that out eventually. The second is we have to make sure that whatever architecture we build can exist in the real world. We don’t want to have our computers only operating at absolute zero and only in one place. We have a lot of work to do to make cubit stable in the real world and that’s mostly a physics and a mechanics problem to try to sort out, and we’re doing things like using lasers to isolate ions and excite them. It’s crazy stuff, but there’s a ton of work going on around, how do you make cubits stable in less esoteric environments so that they can be repeatable and you can scale them?
John Roese: And then the third is, how the heck do you program one of these things because it doesn’t look anything like a traditional computer? And we’re doing a lot of work in the industry to try to figure that out. In fact, today you can go into some of the public clouds and there are companies like Honeywell and others that have quantum computers that you can go and run software against. But it’s only a 20 cubit system or a 50 cubit system and it doesn’t really do much before we can do anything really interesting. And even when we do that, which we estimate is still three, four, five years away best case, and that’s an aggressive roadmaps, we’re not going to replace the generalized compute world. We’re not even going to replace GPU’s and AI processors. We’re still going to have heterogeneous compute. Quantum is just going to be another part of that heterogeneous compute, another accelerator. But it’s going to be a very specialized accelerator that can run some entirely new mathematical functions that are only possible when you have this quantum paradigm in effect.
Kelly Lynch: I want to ask you one more question, and this doesn’t have to do with anything Dell-specific even, but for you, John, what are you excited about in 2021? Personal life, professional life, global scale, what are you excited about?
John Roese: I am particularly excited to see where we end up. I’m an eternal optimist. I think that we’ll have a vaccine. It’ll be rolled out. I think we’ll break the cycle. I think that sometime next year, the world will go back to something. I just don’t know what that something is because we talk as if people have learned our lessons and we understand how to balance the digital and virtual environment and the physical environment and we’re going to end up in a new normal. I think there will be a new normal, but I think human nature says, “We’re going to oscillate back and forth.” We’re going to over-rotate back into the real world and then we’re going to spin back into the virtual world. And I am particularly interested to see how well we navigate that.
John Roese: I’m pretty good at predicting many things, but I can’t predict what normal will be December 31st of next year, and I’m intrigued by it and I’m going to watch it and we’re going to try to influence it because we think the right answer is a balanced relationship between all of this new virtual world that we’ve been able to exist in for a year, and the important things from the physical world, working in balance. My guess is we’re not going to get there in one step. We’re going to rotate all over the place and ended up in all really weird collisions. We’re going to create a lot of confusion and eventually, hopefully, we’ll gravitate and normalize towards an end state that’s something like what I think we should be.
John Roese: But on a personal level, I’m in an observation mode. I really want to see how this plays out. I’m curious to see how humanity rotates back into normal. And I know it will oscillate, but I think it will settle down and I’m going to track it very carefully because for Dell, that’s incredibly important because our job is to be an enabler of human progress. And if human progress is thrashing all over the place, we’ve got to do our best to normalize it and to enable it to calm down. And I think that once COVID is done, the action doesn’t end. In fact, we now have to recalibrate into a new normal, and that usually takes effort and it takes leadership and it takes time. And so, that’s going to be a very exciting thing probably for the second half of next year. But we should reconvene on December 31st of next year and see what normal actually became because I don’t know yet, but I think it could be very interesting.
Kelly Lynch: This has been incredibly educational for me. As you can tell, I have a lot of questions about things that are a little confusing, but I’m excited to see where Dell Technologies ends up, where the world ends up, where we end up, and I’m grateful for your time. Thank you so much, John.
John Roese: Great. Well, thanks for having me. These are always fun discussions to talk about what’s coming next.
Kelly Lynch: Thank you so much for listening this week everyone. I feel insanely lucky to keep producing these podcasts for you and I can’t wait to share more in the new year. Actually, if you have suggestions of topics you’d like, please don’t hesitate to let me know. I mean, my email is just firstname.lastname@example.org. That’s K-E-L-L-Y, the right way to spell it, and I’ll see what I can do. Until next time/until 2021, I’m Kelly Lynch, and this is the Next Horizon, and I’ll hit you on the flip side.