
Sign up to save your podcasts
Or


In this Tech Barometer podcast segment, technology analysts Simon Robinson, Natalya Yezhkova and Steve McDowell discuss the Broadcom acquisition of VMware and how it’s impacting CIOs and IT decision makers.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript (edited for readability):
Simon Robinson: Clearly the market is in flux. It has been in a transitional phase for the past several years.
Natalya Yezhkova: Broadcom is a mature company. It has a certain way of doing business, so VMware culturally business-wise was very different from Broadcom.
Steve McDowell: They’re laser focused on high revenue, high margin business, and every move that we’ve seen them make plays directly toward that.
Jason Lopez: Three technology analysts walked into a room and talked candidly about the Broadcom acquisition of VMware. This is the Tech Barometer podcast. I’m Jason Lopez. The analysts are Simon Robinson of Enterprise Strategy Group who covers infrastructure. Natalya Yezhkova, a research vice president with IDC who writes about enterprise workloads and emerging workloads infrastructure, and Steve McDowell, chief analyst with NAND Research. This was a press conference at .NEXT in Barcelona, Spain in May of 2024. One of the initial questions in the press conference: Why did Broadcom make this move? Natalya Yezhkova began by explaining why companies make acquisitions in the first place. One reason is to grow the new acquisition as a business unit,
Natalya Yezhkova: Or they go the Broadcom way, which I think happens rarely with companies as large as VMware small organizations. When they are acquired, they usually absorb very quickly. Broadcom wants to continue to do business the way Broadcom always did it. They wanted to go into the software business, and that’s what the acquisition of VMware was for them. But they didn’t see a need to adjust to the VMware culture or keep VMware as an independent company or more independent business units. So they want to absorb it quickly into the Broadcom way of doing business.
[Related: Nutanix CEO Paves Way for Those Leaving Broadcom VMware]
Jason Lopez: Simon Robinson said he’s heard no indication Broadcom would integrate or encourage collaboration between VMware and other business units.
Simon Robinson: The focus here is really the opportunity as Broadcom MO (modus operani) was to take a business that has established itself at the core of most data centers globally and enabled quite an incredible change in how data centers are architected over the last 15, 20 years. But I think Broadcom saw the way VMware did that was not efficient. So the capabilities were there, but the efficiency wasn’t, in their view. So that is the core focus. Well, one of the focus areas is to drive more efficiency into how they are structured, how they go to market.
Jason Lopez: Natalya Yezhkova said in terms of hyperconverged infrastructure (HCI), the overall market declined in 2023. Of those that grew, only a few vendors were not associated with vSAN.
Natalya Yezhkova: Customers now approach acquisition of HCI systems because of this uncertainty between Broadcom and VMware. There will be some additional impact on the HCI market. Overall we expect it’ll continue growing, but in the short term, there is some shrink attributed to customers being cautious with purchases of vSAN.
[Related: Wondering What Broadcom Will Do Switches to Mitigating VMware Risk]
Jason Lopez: The centerpiece of the press conference wasn’t necessarily the Q and A with journalists, but these exchanges between Yeshkova, Robinson, and McDowell, starting with the topic of the future lifecycle for VMware products.
Simon Robinson: The market is in flux, and that’s been driven by multiple things. Primarily, I would say from on-premises IT environments to cloud-based IT environments as the major force that we’ve seen over the past several years, and that’s had its own disruptions and impacts on the market. You’ve had the transition caused by the emergence of a whole new class of workloads and applications and the emergence of the technical architectural underpinnings of those things, predominantly driven by the emergence of containers and orchestration all around that. So you kind of got those two major shifts already in flight, and that was causing organizations to make some really big strategic decisions around workload, location, and optimization and building their strategies around that.
[Related: Existing IT Hardware Gets New Path to Hybrid Multicloud]
Natalya Yezhkova: And also to add to what Simon has said, we’ll see the impact across multiple years. So it’s not something that we will see in one quarter and it’s gone. There are customers who up for renewal with VMware. Now there are customers who up for renewal in three years from now, and they still have the old terms of licensing. For them, it’s like a delayed impact and they have time to absorb. But those customers who are currently in the position where they need to make a quick decision whether they want to continue with VMware by Broadcom and adjust to new pricing, definitely in a tougher situation than those who need to make this decision in three years.
Steve McDowell: When you look at VMware technology, what’s interesting about that –– but it’s not true of pretty much anything else except Oracle. Equate VMware and Oracle from this perspective –– it’s that they’re extremely sticky and it’s expensive as hell to change out. And the accountants of Broadcom [did] a really nice job of calculating what that switching cost is. The IT guys, they got to worry about this AI thing. They don’t have time to re-architect how I’m doing my workloads. So they’re pricing it just under the pain threshold for those customers they care about, and they never have to do another piece of innovative engineering ever because these legacy applications, it’s going to be like Oracle. It’s an annuity for VMware forever, mid-range of the market. Man, that’s in turn. I don’t think we know where they’re going to land. We’ve seen some movement back and forth, but those guys, I’d be scared if I were a small company relying on VMware. VMware is not going anywhere.
[Related: Enterprise IT Teams Jump Into AIOps]
Natalya Yezhkova: So I agree that VMware is not going anywhere. We had a range of technologies which went away, but they [are] still here. So mainframe tape… it has been said [often that the] mainframe was dead [and] tape is dead. They’re not. The market shrank significantly, but both still exist. I’m not saying that VMware is going this way, but it’s transforming as a business with this change. But I would agree that it wouldn’t go away. It would look very different like five years from now.
Steve McDowell: That’s a good analogy. If you look at IBM with mainframe, you look at IBM with Power [System Servers]. These are technologies that have not evolved in 30 years. They’re milking a captive user base, and that’s what Broadcom likes to do across almost all of their products.
Jason Lopez: So as one journalist asked, who could take advantage of that position.
Natalya Yezhkova: I can say that from the HCI business perspective, pretty much everyone [can take advantage of this situation]. So with all the uncertainty which exists, I think even companies who worked with vSAN, they start looking for alternative solutions for the HCI offerings. So I think it creates great opportunities for infrastructure vendors. What companies are looking for now [is] integration of resources between dedicated environments and public cloud. They look at security and resiliency. Security and resiliency [are] always on top of the list. They look at modernization of the storage and data management strategies. VMware was in this business as well of providing the tools for managing both dedicated and public cloud resources for investing in security. So they were also in this business, but there are companies outside of VMware who do just what companies are looking for. This whole situation with Broadcom made companies start looking for what else they consider. It’s a great opportunity for vendors to take advantage
Simon Robinson: Just to build on that, on the previous question, again, the dominant motion of the past 10 years has been from on-prem to cloud. In fact, one of the things that VMware talks about is about the risks to going all in on a public cloud. What generally we don’t have in the industry today is portability of workloads, data applications between clouds. That’s still really difficult,
Natalya Yezhkova: But customers want it.
Simon Robinson: But I think, again, this whole trend is going to reinforce [caution for] going all in on anything. [That] has its risks.
Jason Lopez: Steve McDowell followed up by reiterating. He doesn’t think this is fundamentally a technology question.
Steve McDowell: They all smell blood in the water. There’s money to be had, and it’s not a license [but] a switching cost. It’s not going to be a rush to the door because apart from a handful of small businesses, everybody’s on three and five year cycles. There was a rush to renew those licenses ahead of the Broadcom acquisition. So it’s going to unfold quarter over quarter for the next four years.
[Related: Focus Shifts to Migration in Wake of Broadcom’s VMware Acquisition]
Simon Robinson: Well, there’s also the ways in which they’re tied in, right through custom scripting and all of that, depending on the breadth of the adoption on that side. Technology vendors are nothing if not pragmatic. If they see that their customers are asking for something different, then they’re going to respond.
Jason Lopez: Simon Robinson writes about infrastructure for Enterprise Strategy group. Steve McDowell is Chief Analyst at NAND Research. Natalya Yezhkova is Research Vice President with IDC, covering Enterprise and Emerging Workloads Infrastructure. This is the Tech Barometer podcast. I’m Jason Lopez. Thank you for listening. Tech Barometer is produced by The Forecast where you can find more stories on technology and the people in tech. Just go to theforecastbynutanix.com.
In this video interview, Taylor Linton of Hugging Face explains the thought process many IT decision makers go through as they select open source and proprietary AI software to run their businesses.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Taylor Linton: With open source. There’s really two categories of benefits with using open source models, and it’s really broken down into strategic and tactical reasons. To start off on the strategic side, it really comes down to ownership of an ip. Meta has done a great job commoditizing the models and other open source model trainers. So the fact that these companies can take their proprietary data and customize these models, that’s a great opportunity to be really a competitive differentiator from their competition, having their own models that they own that are trained on their proprietary data. Another big reason on the strategic side, it really comes down to cost. And what I mean by that is sometimes your incentives aren’t aligned when you’re using a proprietary model because they charge based off of input data and output data. So the more value you get, the more charged. And so the fact that you can build an open source model and you’re only paying for the commodity hardware, your goal is to get as much usage as possible, whether it’s getting it in front of more customers or getting more AI assistance used from your employees, not only are you able to train your own model and keep it in your environment, but there’s a lot of security benefits.
[Related: Role of Open Source in AI]
Customer trust is huge. Responsible AI is huge, and the fact that you can feel confident that you are sending your data to a model in your environment, it’s a great opportunity to really de-risk it. Another part on the tactical side comes down to latency. If you don’t own the model and you can’t control the infrastructure that it’s running on, there’s many times when requests might take four or five seconds to get a response. So when you have ownership of the open source model, it allows you to make sure that you get that ultra low latency to support your requirements.
So that’s one of the things that people really appreciate with open source models because you can download the model and run it in any environment you want to. So it could be their preferred cloud provider, it could be on-prem, it could be in their Nutanix environment too. But really it comes down to where do they want to run it? Where’s their preferred environment? We often suggest our customers to go grab their own proprietary model, test out the use case, see how it performs. It offers a good benchmark to know roughly where AI is to be able to support the project from an accuracy standpoint. Now, once that POC has been built, that’s where we work with them to say, okay, how should we build this with an open source approach? Now, open source AI is not a silver bullet. There’s really trade-offs too. So when you start looking at cost, these models, they scale very well from a cost standpoint, open source models, and the reasoning is because you’re paying for the commodity hardware that it’s running on.
[Related: Enterprise IT Teams Jump Into AIOps]
The downside to that is if you don’t have a large volume use case, it could require some expensive compute to host that model, and it might not make financial sense to go deploy a model for low usage. In those instances, it totally makes sense to be able to use a closed source proprietary model because you’re only charged for input data and output data. So it can be really affordable in lower volume projects. But of course, as you do scale it out, that’s where it gets very expensive and the scale tips pretty quickly. It’s where it does make sense to deploy an open source model on your own hardware.
Related: IT Team Modifies Open Source NetBox to Help Manage Hybrid Multicloud]
So it’s nearly impossible to try to automate the process of knowing which model to use and the best way to train or adjust the model for a use case. And so our engineers sit down with customers and look at what exactly are we trying to build, what type of data are we going to be sending to this model? What constraints are there, whether it’s cost or latency. And so we sit down with them to understand what is the best open source model for this particular project. And then from there we can sit down and figure out how should we customize this to improve performance, whether it’s speed, latency, or accuracy. These are all trade-offs that really you kind of have to pick one or the other. So we do our best to help them make the best decisions there.
In this video interview, Hugging Face’s Taylor Linton explains how enterprises can use existing open source software instead of building AI applications from scratch.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Taylor Linton: Predating me joining Hugging Face. We started in 2016 to be a chatbot company. That’s where the fun hugging face name and emoji comes from, because it’s supposed to be a chatbot that’s your friend. So we wanted a friendly brand associated with it. But since I joined Hugging Face in 2021, we started out with 15,000 pre-trained models on the hub. And it was so interesting talking to customers then because it was, Hey, there’s too many models out there, I don’t even know which one to start with. And just seeing that grow from 15,000 to 650,000 in a few years, it really shows how much ai, especially open source has grown.
[Related: Enterprise IT Teams Jump Into AIOps]
It gets even more challenging. So since yesterday when you and I both looked at the models, there’s already 2000 more models. It’s crazy. And the tough part is you can’t really automate these decisions on which model to use. It really is context dependent. You need to know what your cost requirements are, what type of input data that you’re going to be sending to the model, because models are trained on different types of data and you need to try to back into the model that was exposed to your type of domain. During the training. It comes down to latency, the particular task. So these are all things that our engineers can sit down with customers and look at what are we trying to build here? What type of constraints are we working with? And we really walk through them, walk through with them the different trade-offs that you make when you’re picking a different model.
[Related: Role of CIO Expands with Enterprise AI]
A lot of CEOs weren’t even aware of what AI was a couple years ago, and then chat GPT came out and CEO’s kids were doing their homework with it and all of a sudden they started hearing about it. So there’s been a lot of interest in companies to try to take advantage of this technology. When folks talk to hugging Face, it’s because we’re the entire open source ecosystem around ai. And so when they want to explore and take advantage of these open source models, they might go to our hub to go grab one of the 650,000 open source models. But also, once they do grab that model and bring it into their environment, we have quite a bit of maybe 20 different libraries or tools that they use to actually go all the way from building to deploying these models. So that’s really when we get involved with folks is to help them take advantage of open source AI.
[Related: Building a GenAI App to Improve Customer Support]
The community around hugging face and just open source AI is incredible. So the fact that someone might release a paper that introduces a new optimization technique, the fact that the entire community can benefit from that and be able to use it on their models, it’s really exciting and it gets folks pretty motivated to be able to introduce a new novel technique and open source it to benefit a larger group of people. I think that open source AI is critical to be able to keep responsible AI across the world. The biggest reason is because the last thing that people want is to only have a few companies that have access to this technology. The fact that people can build in public and you can know certain data sets that were used to train the models. It’s a great opportunity for companies to make sure they are using models that were trained on permissive data. And bias is also a really big area that folks are looking into because of course, human bias exists in these training sets, and that ends up transferring to the model’s bias. So the fact that people can work together and in a community to really look into these data sets and do their best to try to identify and mitigate bias in these models, that’s very important.
[Related: AI and Cloud Native Alchemize the Future of Enterprise IT]
We partner with quite a few companies, but Nutanix has really stood out to me. It really seems progressive to make sure to offer the best tools and technologies are made available to their customers. So what I’ve learned with GPT in a Box, the fact that they’re integrating it with all of hugging faces, open source libraries, they’re making it frictionless for customers to be able to deploy open source models in their environment. You don’t see that often from, and they’re really going a long way to make sure that the best tools are available for their customers.
In this Tech Barometer podcast segment, Justin Mason, associate director of vendor and operations at the University of Canberra, explains how moving to Nutanix software enabled their IT operations to run demanding AI and machine learning courses and keep students engaged in their online studies.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Justin Mason: We are all AHV now. We don’t have any VMware. We’ve been in that place now for a number of years, so we’ve been fortunate enough to miss this whole debacle around VMware and Broadcom at the moment. I know a lot of other organizations are kind of feeling that pressure now, so that’s one less thing we have to worry about.
[Related: Birth of a Hypervisor That Unleashed the Hybrid Multicloud Era]
Jason Lopez: Justin Mason is the associate director of vendor and operations at the University of Canberra in Australia’s Capital City. This is the Tech Barometer podcast. I’m Jason Lopez. What we have here is a press conference from 2024, next held in Barcelona, Spain, as in our previous reporting. These press conferences are a bit different, favoring an informal face-to-face with tech journalists. So Mason started off with some background on how the University of Canberra’s IT organization adopted Nutanix. He oversees the IT strategy for the school, which numbers about 15,000 students and staff. About 10 years ago, they decided to move away from three tier architecture.
Justin Mason: We were early adopters in the sense of when we jumped on the Nutanix platform at the time, we started to look at, well, how can we do this better? How can we do things more cost effective? And that’s when we looked at Nutanix and hyperconverged. Back then, we were primarily a Hyper-V shop, so we started to run Hyper-V on Nutanix. That wasn’t as reliable as we’d hoped, but that was more on the Microsoft V clustering side of things. We then moved over to running VMware on Nutanix as our second POC or proof of concept. That went extremely well.
[Related: Wondering What Broadcom Will Do Switches to Mitigating VMware Risk]
Jason Lopez: Mason and his team decided to fully transition to hyperconverged in 2014,
Justin Mason: And that’s when we decided, okay, we want to go all in on this platform. It was doing what we needed to do as extremely reliable, and that’s when we started to look at it, move over all of our legacy architecture to Nutanix hyperconverged platform. A couple of years later, Nutanix came out with AHV because we were the kind of technology leads at the university. Were always encouraged to try new things, and at the time we were already paying additional licenses and VMware and could save some money in there. That area. It made sense. And fast forward today with Broadcom buying out VMware,
Jason Lopez: Mason explained how that early decision to move to Nutanix allowed the university’s IT efforts to be free of lock-in to a VMware license. Today, the university’s infrastructure hosts hundreds of VMs running on Nutanix clusters.
Justin Mason: We’ve probably about seven different clusters on Nutanix, 50 odd nodes, about 500 workloads, 500 odd VMs spread between the different cluster sets and some of our major corporate systems are running. So the way that things are panned out for the university, we have a cloud adoption framework that’s out of date.
[Related: Existing IT Hardware Gets New Path to Hybrid Multicloud]
Jason Lopez: But Mason says, as their workloads run either on a Nutanix private cloud or in SaaS services or even run a few things on public cloud, they’re revamping their cloud strategy.
Justin Mason: Getting onto public cloud is pretty easy. What we were finding is having skills to cover all the public cloud options. Like having an Azure expert, an AWS expert, a GCP expert wasn’t really financially viable to have that many different skilled resources across all those technology sets. But whereas Nutanix is kind of a single platform that we’ve had for a long time now. So our skill sets are pretty well developed across it. And obviously with a SaaS platform being that it’s a bit of a black box managed by other providers, that works well for us. So Nutanix is definitely going to play a large part in that cloud strategy moving forward.
Jason Lopez: The University of Canberra is a prime example of how antiquated the idea is that it runs anonymously in the background. Mason said the student experience plays a critical role in a university’s reputation, and that experience is significantly affected by the digital platform students use to sign up for classes, submit homework, send and receive messages with teachers and administrators, and to just stay connected to what’s going on on campus,
Justin Mason: Putting in systems now that can track the student experience and kind of flag when students might need additional help. They haven’t been to class in a long time. We have some systems around that like, well, hey, it looks like you haven’t put in your last assignment or something like that. Do you need additional help? So there’s definitely a lot of, and even more of it coming out the digital strategy and putting more importance on that to make sure that the students are at the front of that experience factor.
Jason Lopez: The university’s digital strategy launched earlier this year. It includes 30 initiatives aiming to enhance the student experience while ensuring staff and students develop the skills to use new digital tools effectively.
Justin Mason: It’s kind of one of those things where it’s a 10 year roadmap, years one to three is pretty much well bedded down and spelt out, but those other phases, while there’s a high overview of what they want to achieve, they’re really going to rely on what the outcomes of phase one is that’s going to feed into it. A lot of it is around the student experience, the student journey, and what we’re finding that comes into your question about the digital skills and training for users. I think that’s a really important aspect as well. Obviously, there’s a cloud strategy that’s coming out of the digital transformation. There’s a big piece around integration, especially universities, probably similar for other organizations, so many separate systems and getting them to talk to each other reliably.
[Related: Machine Teaching: How Schools Are Using AI in Education]
Jason Lopez: Then there’s AI. Mason said there are a lot of things coming at it at breakneck speed, or as he stated in the press conference, it’s when everyone’s running around with their hair on fire.
Justin Mason: But from a teaching point of view, it’s not new. We’ve had subjects in artificial intelligent and deep learning and machine learning for quite a number of years, and some of those subjects have been complimented by students using virtual desktops via Nutanix, powered by GPUs to give them the additional resources they need to do those subjects. So not exactly new in the academic world. And by leveraging Nutanix, that’s allowed the students to access to those resources from anywhere at any time. They don’t need to come into a specialized lab running big hulking machines. They can do it from a web browser on their normal machine, whether they’re on campus or at home.
Jason Lopez: He wrapped up with this comment. Organizations considering a shift from VMware to a HV often worry about reliability and skill requirements, but those experienced with VMware find a HV easy to manage due to Nutanix’s user-friendly design.
[Related: Nutanix CEO Paves Way for Those Leaving Broadcom VMware]
Justin Mason: What we found is if you’re good at VMware, then you’ll do AHV in your sleep, and that’s Nutanix credit for making it so easy to manage. So when we move from Hyper-V to a HV, that was a bit of a learning curve for us. But then once you learn it, you kind of learn it. We found them when we moved to a HV when we were doing the migration, and part of that migration was outsourced to Wipro. There’s not much pushback here, and they’re doing it really quick, and it was kind of like, well, I think they were seeing the benefits of moving to it. What about getting external resources in that you might need to help with Nutanix? We’ve never had to do that. We’ve never had to go looking. There’ll be times in the university where we’re doing certain projects or certain initiatives. We need to bring in a resource to do something specifically like a T Net expert or a Microsoft Dynamics expert. But we’ve never found ourself having to bring in a Nutanix expert, so to speak, because I guess that speaks for itself. It’s pretty easy to manage.
Jason Lopez: Justin Mason is the associate director of vendor and operations at the University of Canberra, located in Canberra, Australia. His press conference was recorded at 2024 dot-NEXT, which was held in Barcelona, Spain. This is the Tech Barometer podcast. I’m Jason Lopez. Tech Barometer is a production of The Forecast. It’s at the forecast by nutanix.com where we’ve got technology articles, podcasts, and video for you to enjoy.
That’s theforecastbynutanix.com. Thanks for listening.
In this Tech Barometer podcast, Northeastern University professor Dr. Norman Jacknis explains why IT leaders must concentrate on delivering business value while understanding essential and emerging innovations such as artificial intelligence and machine learning.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Dr. Norman Jacknis: Part of the reason why you need to get started is you are learning how to do this stuff well. So you have to get started. And if you don’t get started, you’re going to fall behind. Right now, one of the reasons why companies that are using this are ahead of their competitors is they started a mile back. They’ve made the mistakes. They’ve figured out what works and what doesn’t work. And they’re running. And you’re in a race. If unfortunately your competitors who are ahead of you keep on running as fast as they are and you can’t run faster, you’re never going to catch up to them. I think also it’s important just for your own jobs because in most companies, it’s at least crossed the mind of the CIO to ask the CIO about how to use AI.
[Related: The CIO Mindset for Embracing AI for Enterprises]
Jason Lopez: Norm Jacknis teaches mid-career technologists how to be CIOs. He’s Professor of Practice, Innovation and Entrepreneurship at Northeastern University in Boston, Massachusetts. In his career, he was an executive at Cisco and the CIO of Westchester County in New York. He was also chairman of a chapter of the Society for Information Management, a prominent organization for IT leaders and CIOs. The role of the CIO has changed dramatically. He says, it used to be the CIO just made sure all the technical parts worked. Computers stayed connected. Email wasn’t lost. Software was correctly installed. As he refers to it, making the trains run on time. But then came the cloud. Here’s an insight he gives his students.
Dr. Norman Jacknis: I said, you need to ask yourself, what am I doing all day? Because the stuff I used to worry about, like the data center, big data center being up all the time, I’ve outsourced a lot of that stuff to the cloud provider. So they’re worrying about it, not me. A lot of software I’m getting is sort of off the shelf and probably most of it is in the cloud as well. And again, somebody else is worrying about it. So what do I do? Sit here all day and twiddle my thumbs? Well, everybody finds ways of filling the day. But I said, why don’t you think about what you can really do in your organization? Think strategically. So now your role is not necessarily just to have the trains run on time. You still have to worry about that. But it’s more, how can I, as a technology leader, make this stuff strategically useful for my company?
[Related: How Nutanix Built a GenAI App to Improve Customer Service]
Jason Lopez: Today, a CIO has to think not just about moving data around, but what that data means. And with the rise of machine learning and AI, it’s become critical that CIOs take this on.
Dr. Norman Jacknis: The leader of this really should be the CIO. Unfortunately, in a lot of organizations, that’s not the case. Sometimes it’s been led by chief marketing officers and in other cases by a chief innovation officer. But when that happens, that’s a poor reflection on the CIO. The CIO has not done his or her job properly, if that’s the case.
[Related: Seeing AI’s Impact on Enterprises]
Jason Lopez: But here’s the kicker in what Jackness is saying. A CIO who’s the point person in an organization’s development of AI isn’t just doing a software migration in the background, but is actively part of the business strategy.
Dr. Norman Jacknis: I always used to joke that the IT staff always used to go around telling everybody they had to change, they had to adopt this new system, this new technology. And if you looked at their own behavior, they were always the last people to adopt change themselves. But yeah, that’s part of your job. You are a change agent. You’re a change leader. Particularly when you’re talking about something like artificial intelligence, which is threatening to some people. Think about it. If you’ve got some use of artificial intelligence that identifies where there might be some new potential customers, and you also even have a way of phrasing the marketing material that would go out, and you present this to the chief marketing officer, sometimes you’re going to get a reaction that’s, I’ve been in marketing for 40 years and you think you know what? And so there’s a significant change. This is worse than bringing in, when you brought in, for example, an SAP system, it was aggravating to people because they had to change procedures, but you weren’t threatening their professional identity. You can do that with artificial intelligence. Now you have a really strategic technology. You have the opportunity to think about how it can be strategic for your organization. That’s going to get you the invitation to sit down with a CEO and other C-suite colleagues. We go through the major algorithms, but the focus is on how you will use it.
[Related: IT Leaders Get AI Ready and Go!]
Jason Lopez: Jackness says the course materials reflect the real world. There is work focused on AI, data centers, and cybersecurity, but also that real world component. Again, students are in their mid-careers. So each semester they have to defend their ideas before a panel of industry experts, as if they were going before a capital investment committee or VCs. The goal is to empower them to be able to implement things and put together teams.
Dr. Norman Jacknis: How can I get this off the ground without asking the CEO for $15 million? And for that matter, we even go over the issues of change management. Okay, so you’ve done this thing. How do you figure out which groups you start with in your company? How do you get it deployed? So we’re really helping them very much to sort of think about this as a leader, in addition to what the technical issues are. The later semester as well, we deal with operational issues. Okay, you’ve got this great idea of yours and your thesis approved. What are the day-to-day issues that you have to deal with, including, for example, ethical issues, as well as operational issues, dealing with staff, stuff like that. Think about how a CIO spends his or her day. Over the course of a year, we want to make sure we touch on all those various questions that cross that person’s mind and bring the best knowledge we have, the most recent research about what works and doesn’t work, and we want to bring that to bear and actually have them apply it.
[Related: AI in the C-Suite: Using Artificial Intelligence to Shape Business Strategy]
Jason Lopez: Aside from the basic tech knowledge, which is baked into the curriculum, he brings his experience into play by emphasizing what’s beyond the classroom.
Dr. Norman Jacknis: I think they need to keep up with the world, you know, so they should spend at least some of their time just keeping abreast of developments in artificial intelligence, because their colleagues are looking to them as the technology leader, right? You’re the CIO. You’re supposed to know more about technology, including artificial intelligence, than the rest of us do, right?
[Related: Enterprise IT Teams Jump Into AIOps]
Jason Lopez: And one of the advantages of today’s computer science, you don’t have to spend a lot of money to get experience, just a laptop and a connection.
Dr. Norman Jacknis: I mean, there’s even software now that will provide a lot of the algorithms for free, open source software. There’s a lot of open data. The efficiency of the algorithms has really improved. So it’s not going to use up as much of the cloud resources as you might have in the past. Some of the stuff you can even start out doing on your own laptop. And then when you go to production, you can move it to the bigger data centers in the cloud. But at least you can get a start. You can be able to show somebody, hey, I can do a good job of identifying what customers might be highly probable of churning. So we can focus in on maybe doing something extra with them.
[Related: The Amalgamation of AI and Hybrid Cloud]
Jason Lopez: The program at Northeastern awards a Master’s of Science, but Jackness points out it’s learning by doing. It emphasizes leadership skills in a technology environment. Perhaps think of it this way. What can you learn to bring to a company to help it succeed?
Dr. Norman Jacknis: Well, I’m not very academic, I suppose. So it’s been easy for us because basically what I do is teach some of the theoretical stuff and then also have them apply it to their jobs. I teach a course on artificial intelligence and machine learning for technology leaders. I go through the major algorithms so they understand what they’re supposed to do. But their assignments are not to do heavy duty data science work, but instead to figure out, OK, strategically, how is this going to apply to my company? And at the end of the semester, they’re actually writing a memo to their CEO on what the AI machine learning initiative ought to be for their company. So it’s very much a marriage of these two sides. And actually, it’s one of the reasons why I moved to Northeastern, because Northeastern is known for what’s called experiential learning, which is basically a philosophy of education that you will learn best when the concepts that you learn are actually applied to your life, to your work.
Jason Lopez: Norm Jacknis is Professor of Practice, Innovation and Entrepreneurship at Northeastern University in Boston, Massachusetts. This is the Tech Barometer Podcast. I’m Jason Lopez. Thank you for listening. We also have another podcast with Professor Jacknis, and that focused on the role of CIOs implementing AI in their organizations. The Tech Barometer Podcast is produced by The Forecast. You can find more stories on technology and people in tech at theforecastbynutanix.com. That’s theforecastbynutanix.com.
In this Tech Barometer podcast, Dr. Norman Jacknis explains how CIOs can adapt to rapid change driven by AI and the value it can bring to organizations.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Dr. Norman Jacknis: If I ask you to travel from New York to Boston, you pretty much know what your options are. You can fly, you can take a train, you can drive, you can get a map from Google, whatever. There’s a whole bunch of things. It’s pretty straightforward, and you might run into a traffic jam or something like that. It’ll sell you down, but basically you know where you’re going and what it’s going to be like when you get there. With ai, particularly generative AI these days, it’s more like the Louisiana purchase in the Lewis and Clark Expedition. You’ve got this whole big territory that you have to explore and you’re not going to know what its value is. I mean, ultimately in the long run, it’s the best $3 million investment the United States government ever made, but you don’t know when it’s going to start delivering. You don’t know exactly why it might deliver. The whole nature of artificial intelligence is it’s an iteration, it’s an exploration, and so that’s one of the reasons why the people who expect to see a return in the next quarter or the next two quarters are really missing. The point about what this is about.
[Related: AI and Cloud Native Innovation Spark Explosion of Business Apps]
Jason Lopez: Norm Jacknis is Professor of Practice Innovation and Entrepreneurship at Northeastern University, which is located in Boston, Massachusetts. Before coming to Northeastern to teach technology leadership, he was an executive at Cisco and the CIO of Westchester County in New York. Those are just a couple of his consequential stops throughout his career. And in this podcast, you’re going to hear about the role of the CIO in establishing AI and organizations. What makes this compelling is that as a professor, he’s had to gather together his deep experience in technology and organize it to teach future CIOs, and we gleaned some of his best insights. This is the Tech Barometer podcast. I’m Jason Lopez, so let’s pick up from where we started this podcast with his comment about understanding what AI in the enterprise is.
[Related: IT Leaders Get AI Ready and Go!]
Dr. Norman Jacknis: There was some hype clearly, but I think there’s been pushback, particularly from CEOs, they see the value in this stuff these days. What those people are talking about is generative ai, the more traditional machine learning for prediction, for improving interactions with customers so forth. That’s already a given. You can see companies that have been using that and have been making money from that, but even on the generative AI stuff, I think they hyped it up too much, and I think outsiders who haven’t done this successfully don’t understand the nature of working with ai. It’s different from a traditional software project.
Jason Lopez: Jacknis says the hype around generative AI has put so much attention on language processing. Perhaps it’s easy to forget about machine learning, something he reminds students who are mid-career executives.
Dr. Norman Jacknis: There’s still a lot of low hanging fruit, if you will, more traditional machine learning. That stuff can be used to help identify potential customers, help identify problems ahead of time that you can fix all that sort of stuff. So that kind of thing is still very useful and can be done, and the generative AI stuff properly used also can do some things to make it easier for your users and customers to have access to the computing resources they need instead of having to be subject to the constrictions. The restrictions that have come up about over decades of computer science history
[Related: How Nutanix Built a GenAI App to Improve Customer Service]
Jason Lopez: Those decades ago before it was AI or generative AI or machine learning, it was called computer automation where for example, you took what was on paper and put it on a screen.
Dr. Norman Jacknis: Yeah, no, I suppose it was an improvement. I’m not sure how much it was annoying, and of course, they’d ask the same questions over and over again the way you did on paper forms. Now with generative ai, one of its advantages is it can be much more conversational and it can be smarter about what to ask. So if I put down than I’m mail, it’s not going to ask me how many months pregnant I am and things like that. It’s a lot smarter and it makes that whole human computer interface a lot easier.
Jason Lopez: He talks about AI as a strategic tool. There are ways to deploy it that aren’t so great and ways that are successful.
Dr. Norman Jacknis: One of the things about the successful uses of AI is that they help break down the silos in companies because you have to, the folks who are expert on the algorithms don’t know the business. For this to work, they need to speak to the business, and the nice thing about it is AI is sexy enough. Everybody’s quite willing to work together on this. So I think that’s actually kind of helpful. So there’s a strategic value in this as well.
[Related: AI in the C-Suite: Using Artificial Intelligence to Shape Business Strategy]
Jason Lopez: Jacknis gave us an insight in what he teaches students in his technology leadership courses, how a business can get started in ai.
Dr. Norman Jacknis: Don’t go out and hire a whole bunch of data scientists and programmers to write your own AI cloud providers, as you know, offer most of the various options for AI tools. There also is open source stuff. One of the things I use with some of my students is a product called nine K-N-I-M-E, which is from Europe, open source and free. Got a lot of these algorithms there. But start off with that. So the focus is on how do I want to get the data and present it to the computer so it can come up with the best predictions and not worry about getting stuck writing all kinds of blower level code.
Jason Lopez: Once a CIO’s team has become successful in AI deployment in the cloud, measuring that by an uptick in revenue, Jacknis says about 5%. Perhaps it might be time to swap out some capabilities for something more proprietary, but he emphasizes use the cloud and operate on laptops before doing anything on-prem.
Dr. Norman Jacknis: When this started out maybe a few years ago, a lot of the people in AI and machine learning wanted to use the internal data centers that their companies had because there was a big concern about data being proprietary. They didn’t want a lot of people know it and all that sort of stuff. And then the CIOs who were not part of this process pushed back and said, you want to run this program, which is going to eat up gobs of computing resources while I’m trying to run a payroll or a CRM. No way. So in some respects, they were told internally, push this outside, put this in the cloud because we’re not going to let you do this internally.
[Related: Seeing AI’s Impact on Enterprises]
Jason Lopez: And another piece of advice from Professor Jacknis, create the culture around AI deployment. He’s seen failure around treating AI in the enterprise as a technology project, treat it as a business project
Dr. Norman Jacknis: As opposed to we have a set of tools that will help us learn more about the environment in which we’re operating in, help us identify customers better, all that kind of stuff. Instead of saying, Hey, this is a joint effort of everybody who’s got some experience worthwhile in our company to work together. They treat it as a technology thing and then it fails partly because typically what they’ll do is they’ll think, oh, I have a whole bunch of data. Lemme throw it at the computer and see what happens. Oh yeah, the computers are good, but not that good.
Jason Lopez: He goes on to say that when establishing an AI component to business operations, make sure you bring in some people who truly get the business side of it. When he spoke about this, he referred to research showing just several years ago that new AI teams were made up of mostly data scientists.
[Related: Making IT Infrastructure AI Ready]
Dr. Norman Jacknis: Ninety percent of the people in their AI projects were all data scientists. A couple of years later, it was maybe a quarter or 20% because what they found was they needed people who understood the business side at the beginning of this process, and they needed people who were better at knowing how to implement systems. Once you figured out what you have and the actual amount of data science work was small, the computer’s pretty good at this. In fact, you’ve got some computer systems now that will sort of basically be automated data scientists. What they can’t automate is the business understanding that you get from human beings, and they obviously are not going to be able to fully handle the implementation with human beings to make this stuff come to life in your company. This field of AI is moving very fast, and every problem you read about, there’s a bunch of smart people working on fixing it. Sooner or later they will. In fact, there’s a story that came out written by a CIO and he said Everything we thought was hard turned out to be easy. Referring to the conversational ability that you have with Gen ai, which was something the year before it came out, people would’ve said, oh yeah, we’ll get that 2030, maybe 2035. And here we are.
Jason Lopez: Norm Jacknis is Professor of Practice Innovation and Entrepreneurship at Northeastern University. This is the Tech Barometer podcast. I’m Jason Lopez. We have another podcast in this series featuring Professor Jacknis and in that one, and check it out on forecast by nutanix.com. Check out the changing role of the CIO. Again, this is the Tech Barometer Podcast. You can find more stories on technology and the people in tech at the forecast by nutanix.com. That’s all one word, the forecast by nutanix.com.
In this video interview, Nutanix cloud native technology expert Dan Ciruli describes the trends and technologies powering an explosion in new applications, particularly those with AI capabilities.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Dan Ciruli: The more we realize the benefits, apps can do this, apps can do that, and the more use cases that over time we have more data, we’re gathering more video, audio data, they just enable more things. And so, I have no doubt that as we, you know, data centers become more powerful, the things in our pockets become more powerful, the computers at the edge become more powerful, people have more and more ideas, we’ll continue to write more applications than ever, you know, ten years from now we’ll be blown away at where, you know, container-based applications are running and what those things are accomplishing, and all of that means, yes, more apps.
[Related: AI and Cloud Native Spark Explosion of New Apps]
I’m not the first person to make this analogy, but watching AI right now reminds me of late 90s, mid to late 90s internet, and just kind of realizing we don’t know all the ways this technology is going to change everything, but it’s clear it’s going to change everything, right? And that’s where AI is right now. It is, you know, let’s say it’s mid 90s, because it’s in that time when you just knew it was going to be incredible, e-commerce hadn’t been invented yet, I mean, companies really weren’t even figuring out how to market on the internet yet, they were barely putting up websites and taking out, you know, radio ads to tell you to go visit their website, but you just knew it was going to transform, it’s going to transform technology and in turn society. AI is at that, you know, it’s very early on, but you can already tell there is no doubt, and that in itself will mean lots more apps, right, and AI will be, it’ll be in the data center, it’ll be in the cloud, it’ll be at the edge, it’ll be on your phone, it’ll be on your watch, it’s going to be everywhere, and all of that will be transformative.
[Related: AI, Cloud Native and Hybrid Cloud Fuse to Run Apps and Data Anywhere]
Virtually all apps that are being written today are running on cloud native, that’s true of all the AI apps, right? All of the big companies, they’re running all of that AI stuff on Kubernetes, for sure, you can’t, I don’t know of anybody who’s really writing new AI based apps, running those models in VMs, that’s all happening in Kubernetes. So in that sense, AI is built on cloud native and is an example of, it’s a new application, by the way, and it’s one that companies are starting to run in the cloud and just like their other workloads, they’re going to want to run on prem, for sure. Your provisioning timelines are different, you’ve got to buy that hardware and it’s expensive, however, that’s where your data is, you don’t want to send all of your most important data that you need to train models on up into the cloud, so it’s another one of those. On the other hand, in a sense they’re orthogonal too, because ultimately when you’re consuming an application, you don’t care if it’s running on cloud native, you don’t know if it’s running on cloud native, when I open up my bank app and I check my balance, that might be running Kubernetes on the background, I don’t know, right? Eventually it might be using AI in the background, I don’t necessarily know. So in a sense they’re orthogonal. One other thing is I think that when we talk about gen AI, it is one of the things that is going to solve the usability problems. We hear a lot about skills gaps and cloud native, and I can’t hire people who understand Kubernetes well enough to run it and then be asked to do that. To me that’s not a skills gap, it’s a technology gap. The problem there is that Kubernetes is too hard to use, and AI is going to be one of the things that solves that. And when we get to what we call intelligent infrastructure, or invisible infrastructure, infrastructure that is using ML to understand, oh, here’s a problem that maybe hasn’t happened yet, but is about to happen, and here’s what we do to solve it. And you can imagine a day, and we’re already seeing this, we’ve already got AI chatbots built into our software that allow you to say, in plain English, why am I seeing this crash loop error, and what can I do about it? And have AI help to solve that problem. So AI will help us bridge that skills gap by making the technology easier to use.
[Related: Making IT Infrastructure AI Ready]
Almost all application development now is happening in containers, in Kubernetes. What happened over the last 10 years, Kubernetes is having its 10 year anniversary, but the first three or four years was very experimental. So more over the last five years was, yes, organizations were adopting it, but they were adopting it mostly in the cloud. Because the cloud vendors, Google, where I used to work, and Amazon, and Azure, were all making it very easy to get a Kubernetes cluster. And so for many organizations, for the last five years, new development efforts were happening in the cloud, and they were happening in Kubernetes, everything on-prem was still running in VM. There’s two things that are happening. One is that now even some of the stuff that’s being written and run on-prem, they’re saying, well, this is new development now, we want to do it in this modern way. And two, companies are rethinking the economics of cloud computing, and realizing that essentially renting is more expensive than buying, and doing all of your stuff in the cloud is very expensive. So they’re saying, hey, we like these benefits of running in the cloud, we like this idea of velocity, we can get changes into production faster, but we need to be able to do that on-prem too. And so increasingly, IT departments are being asked, hey, how can you give us Kubernetes clusters? You’re the one who can give us VMs on-prem, how can you give us Kubernetes on-prem?
[Related: Building a GenAI App to Improve Customer Support]
Cloud Native started, I’ll say it started at Google, although that’s not entirely fair. There was a bunch of people at a bunch of different companies that were trying to figure out how to ship software faster, how to make it from the time someone had an idea, put that in code, put that into production, could reduce the amount of time to do that. And at the same time, also make sure that when you did, it would be scalable enough to use at some of these big companies. And Cloud Native is what we use now to describe what they were developing back then, which was a, it’s essentially, it’s more layers of abstraction between the developer and the machine that the code is running on. Because traditionally, getting a code to a piece of machine involved server installs, and server installs are notoriously complicated and complex, and it would take a long time to get something just installed and running. And with Cloud Native-style architectures, you could put together a pipeline where a developer wrote some code, they checked it into source code, pushed it to the source code repository, and then it kind of automatically and automagically got deployed to a server somewhere, no one cares where, no one’s ever going to SSH into that server. So Cloud, and then over time, a bunch of this technology got open sourced or created again in open source, and now when we refer to Cloud Native, we refer to these projects like Kubernetes, that ultimately are aimed at this, and aimed at how can we make it faster for developers to get code into production.
[Related: AI and Cloud Native Alchemize the Future of Enterprise IT]
In the beginning of cloud computing, we would talk about hybrid, we would talk about hybrid as a temporary state, and we really thought that hybrid was going to be how people ran until all of their stuff was in the cloud. And I think now where we are as an industry has realized that hybrid is not a temporary state, it’s not a transitional state, this is how most organizations will be running for the next decade, right, for the rest of my career probably, is that, oh no, we will be absolutely using the cloud appropriately, but we are still, we’re investing in data centers, we’re realizing for reasons of cost efficiency, for sometimes data sovereignty, sometimes just control, there’s stuff we’re going to want to run here. And so solving hybrid, not as how do we have this temporary bridge, but how do we put in place something that permanently lets us use the cloud and the data center appropriately, taking into account costs, taking into account where the data is, and taking into account sovereignty rules. So that’s what’s different about hybrid and multi-cloud today, is that companies are realizing this is a permanent state, and how do we make sure that we’re putting in place the right structure so that we can take advantage of each of them for what they’re good for.
In this video interview, NAND Research Chief Analyst Steve McDowell explains how IT decision makers assess strategies and infrastructure needed to run artificial intelligence capabilities.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Steve McDowell: I think we’re all getting AI fatigue. Every product briefing I go to, if it doesn’t have the word AI in it, I’m in the wrong room. That’s all anybody wants to, whether it’s relevant or not. And I think a lot of these technology transitions, whether it’s the internet, whether it’s the smartphone, whether it’s the PC, going way back, there’s a brief burst of time where we’re very excited about the technology pieces, but the value of technology comes in how I use it. So have we hit a wall on AI? No, but I think it’s time to change the conversation. Let’s stop talking about GPUs. We’ll stop talking about who the providers are, and let’s start talking about what are we going to do with it and how’s it going to change my business? Because that’s the interesting conversation. The way I think about AI and we talk about AI, now I’m going to talk about generative AI, but I’ll just call it AI. What’s driving the current moment? I mean, we’ve had AI for a decade, kind of the modern form of AI for a decade. We use it. We use it for predictive analytics, image recognition, retail, whatever. But what’s really impacting what’s about to impact enterprise is generative AI.
We’re seeing a couple of things, and I think 2023 was really the year we figured out how to make these models. And right now we’re going through a kind of rapid phase of how do we make it safe for enterprises? And I look forward over the next 18, 24, even 36 months, and it’s really how do we deploy that in the enterprise from an IT perspective? That means a couple of things. One, I need to pick a partner who’s going to be my generative AI provider. I have a core set of functionality. There’s only a handful of companies in the world that could train these models. It’s OpenAI, it’s meta with Llama, it’s Anthropic, just a handful of these companies. And the way it deploys in the enterprise is I take this large language model and I fine tune it with my own data.
So if I’m an IT guy right now, I’m going down the path of how do I, I’m just trying to deploy. I’m trying to enable my enterprise to use it. And then when it comes to using it, there’s really two pieces. There’s the piece that’s very business focused. How am I going to use AI to enable the next or the next iteration of digital transformation? It’s going to change all our lives, but it’s also, how do I use it to make my own IT operations more efficient? I’ve been to 15 conferences this year, and the theme for 2024 is AIOps, right? Even Nutanix is announcing capabilities around AIOps.
[Related: Pivot Past the Enterprise AI and Cloud Native Hype]
Do I trust it? How do I trust it, how do I deploy it? Where do I use it? A lot of decisions are happening for the poor enterprise IT architect. So what’s happening with ai, we’re trying to figure it out. We’re at the phase where we’ve invented the technology right now, we’re in the enablement phase, then we’re going to start to use it, and then it’s going to drive the transformation. And I look at every kind of big technology, every transformational technology kind of follows this path. Although the timelines are getting much shorter, right? It’s understand it, it’s enable it, it’s play with it, and then it’s going to drive change.
It doesn’t have to be expensive to get into AI as a user and consumer of the technology. And I think where it’s disrupting the tech industry and driving and forcing a lot of this conversation about the technology is it’s very expensive. It’s very complex. It doesn’t look like anything I’ve touched before as an IT guy, or it looks like scientific computing. The cloud guys are solving this for me. They’re managing the infrastructure. They’re buying these expensive GPUs. They’re amortizing the cost over multiple users. Things that I don’t have a budget or capability to do as an enterprise, where that’s causing disruption in the industry is, well, if AI is driving the industry right now and cloud is taking all those dollars, and my company’s name is Dell or HPE or Lenovo, where am I getting my revenue? I mean, the server market was already down. And if those dollars now are being prioritized to cloud, that creates a real dilemma.
[Related: Cloud Vendor Shakeup Puts Focus on IT Resilience]
If I’m looking at this as an IT practitioner saying, where’s the value of AI to my enterprise? I don’t care about hardware. That’s why I like GPT-in-a-Box that the Nutanix is delivering because that’s a software set of capabilities. I can deploy that if that makes sense, right? I can deploy that at the edge, if that makes sense, or I can roll that out in the cloud, if that makes sense. So I think a lot of tech companies are trying to prove their relevance around AI. And I’m not saying they’re irrelevant. It’s just causing a lot of disruption. It’s going to change the way that we think about infrastructure.
There are two pieces of AI. There’s training and there’s inference, right? As a business user, the value is on the inference. An inference is when I take an AI model and I throw some of my own data against it, and it gives me back results, the speed, the time to value, the time to decision is the AI. I say, the closer to your data that the AI is, the faster I’m going to get time to value. And where we look at where, and this is not even a generative AI thing so much when we talk about AI, the most prevalent use of AI is image processing, whether it’s for manufacturing, whether it’s for retail, whether it’s for automotive. If you have a car made in the last 10 years, you have so many sensors in your car, it’s not efficient to take the cameras in your car or the cameras at the seven 11, send those up to the cloud to be processed and send them back down.
[Related: Living Workflows of AI at the Edge]
If I’m running a retail establishment, I might not have internet. It may go down, there may be a storm. I can’t shut my business down when I lose the internet. So by moving those inference functions to the edge, I get all the value of that AI where it makes the most sense. And there’s a couple of things that came together to make this the moment in time where that happens. One is we’ve been talking about the value 5G is going to bring to the world. I get all of this high bandwidth wireless everywhere. We rolled that out without really a killer application. And then we started all this generative AI stuff, and we got really good at AI. And now part of what that did was all the focus on high-end ai, kind of the natural curve of technology is all of the AI inference capabilities required for mid-range, low level ai, dirt sheet, dirt sheet, right? The camera you recorded me on right now can probably track my face. It’s doing inference and facial recognition on a $30 processor. It’s AI at the edge, it’s practical, and it brings true business value. Now, where it gets complicated when we talk about edge, edge is anything outside of the data center where there’s no IT guy, let’s call it that.
[Related: Building a GenAI App to Improve Customer Support]
It’s a couple of edge segments. There’s robo, remote office branch office. To me, those are extensions of your data center. They’re well controlled environments where it becomes really interesting or kind of mass deployments, whether it’s a convenience store, whether it’s factories, whether it’s smart cities, and I have sensors on the lights for traffic control. Managing all of those becomes a challenge. So I’m deploying AI at the edge for these kinds of use cases, but I also have to manage security patches, updates. I got to push new models down. Sometimes I got to bring data back. So what’s the framework look like for that? Well, guess what it, it’s not that different from how I’m doing hybrid multicloud. The boxes are just a lot smaller. So we did this paper and it’s like we’re driving AI to the edge for all the reasons I just said. But then it was really more about, okay, we’re pushing AI to the edge. How do we manage it and how do we make it efficient to manage? And that requires, again, it looks a lot like multicloud, but it may be a more unconstrained environment.
In this Tech Barometer podcast segment, Nutanix President and CEO Rajiv Ramaswami talks to tech reporters about why IT leaders turn to Nutanix to counter changes they face after Broadcom acquired VMware.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcript:
Jason Lopez: In this story, you’re going to hear the voice of Nutanix CEO, Rajiv Ramaswami recorded in a conversation with journalists. This was not the typical press briefing of the speaker at a podium with prepared remarks, but rather he sat at a table with tech reporters in an open-ended back and forth. This is the Tech Barometer podcast. I’m Jason Lopez. The journalists led the discussion, which zoomed in on Nutanix’s view of Broadcom’s acquisition of VMware.
Rajiv Ramaswami: They acquire companies that are fairly mature, that they have good market share, that have good products and fairly sticky ones, and then they do what they need to do to maximize the profitability of those companies over the next three years. And their model is one of continuous acquisition. Acquire, optimize, acquire, optimize, right? Because then that’s the model. So you keep acquiring, you keep optimizing. So they’ve done well with that model so far. They’ve done very well for their shareholders.
[Related: Cloud Vendor Shakeup Puts Focus on IT Resilience]
Jason Lopez: Companies go through stages. Ramaswami pointed out such as an incubation period, then scaling, then maturity. He said VMware had arrived at that stage.
Rajiv Ramaswami: Their growth was slowing. They had kind of largely saturated their markets that they were playing in. vSphere had what, 80% plus share in the market? That’s the stage. That’s when you start thinking about optimization. You cannot grow anymore. So if you were to take us and you do the same squeezing what we got, 25,000 customers, we still have huge amount of opportunity to grow and gain share.
[Related: Slew of Changes Drive VMware Customers to Consider Alternatives]
Jason Lopez: For comparison. Ramaswami explained that Nutanix wouldn’t be a good target for an optimization strategy. It still has a high ceiling for growth, meaning it can essentially double its customer base.
Rajiv Ramaswami: That’s what we are out to do. So we are still in the scale phase and we’ve told our investors, we are profitable, we are growing. So it’s a different status.
Jason Lopez: One of the things Nutanix has been able to do and continues to do is to adapt to the market. The company started with a hardware focus and has shifted to software infrastructure,
Rajiv Ramaswami: But that’s not where we want to be tomorrow. But if you look at the future, the future is around being a platform for modern applications. And that’s why we bought D2iQ, for example. We are investing and expanding our platform offerings. We are investing in cloud native AOS for containers. I think as a company we are in a better place to evolve our customers to the next stage building and running modern applications on these platforms.
[Related: AI, Cloud Native and Hybrid Cloud Fuse to Run Apps and Data Anywhere]
Jason Lopez: As the conversation continued, Ramaswami began talking about the realities of customers making the switch from VMware to Nutanix. The bottom line is the rise of VMware pricing.
Rajiv Ramaswami: The vast majority of VMware customers today are running on vSphere only, and those customers mostly are not perpetually (licensed). And so those customers are not being forced to buy just a subscription, but the full bundle. Customers have to buy the full VCF stack.
[Related: Existing IT Hardware Gets New Path to Hybrid Multicloud]
Jason Lopez: In addition to the issue of being locked into a subscription model with rising costs. There are other questions.
Rajiv Ramaswami: What’s the level of innovation going to be? What’s the level of support going to be like if the model is to double the ebitda, the profits within three years, which is what they’ve said publicly, what would happen there?
Jason Lopez: Another aspect of switching the VMware environment versus the Nutanix environment. An engineer who has worked with VMware tools for 20 years will have to relearn some new ways of doing things. But Ramaswami pointed out, Nutanix is built for adoption. Its processes are easier to learn.
Rajiv Ramaswami: It’s not a huge learning curve, but it is not exactly the same. We do a lot of things differently. In some ways, many of them simpler, and we also have a lot more focus on the data side of the equation compared to VMware.
Jason Lopez: And finally, Ramaswami addressed the question of the VMware stack of the hypervisor connected to three-tier storage.
Rajiv Ramaswami: In the past when people chose Nutanix, they had to adopt hyperconverged (infrastructure or HCI). And a lot of VMware is not hyperconverged. They’re hyperconverged, but the big chunk of their rest estate is really three-tier storage, right? Hypervisor connected, three-tier storage. And so that conversion, we were doing it at a certain pace in terms of converting that three-tier, right? HCI. That’s how we grew as a company. That’s what we did. And now we have an opportunity to go back in and also potentially just replace a hypervisor and get a faster entry. We have seen a huge amount of engagement increase in customers, so lots more customers concerned and upset and talking to us. Many of the customers actually signed multi-year deals with VMware before the acquisition closed. And so that gives them three to five years to do whatever they need to do. And so they’re not in a rush to necessarily do anything at this point.
[Related: Seeing AI’s Impact on Enterprises]
Yes, a lot of them are kicking the tires. They’re testing our product, but to go from there to a sale and actual migration takes time. So they’re not in a rush. The second factor is the hardware, right? At what point in time have they invested in hardware? When does it come for refresh? Because that’s the time at which you can actually make the migration to hyper converge because it’s an architectural choice. And the third point, I’ll tell you, there is much more awareness of what would happen, but it depends on the country and where you’re in. For example, I was in Japan recently, and a lot of Japanese customers were just hoping that everything would be fine because they believe in these long-term relationships and they believe and trust them. And now they’ve seen what’s happening and now they’re concerned. But it took them a year, and they did nothing. Right? And now they’re starting to think about, okay, maybe I should have a plan B. So it takes time. Have we seen a big spike in our business? Immediate business as not yet? We’ve brought on investors that this is going to be a multi-year journey.
[Related: AI and Cloud Native Alchemize the Future of Enterprise IT]
Jason Lopez: The core of Ramasswami’s message is that Nutanix and VMware within Broadcom, though they compete in the same space, have different agendas. Nutanix is aimed at growth. Broadcom is aimed at billing an existing customer base.
Rajiv Ramaswami: That’s their model. So they’re actually going down a true and trusted strategy. Typically, private equity companies do this, but they’re doing this in the public market and doing that very well, I should say, executing on their strategy. I just want to say that’s not our strategy. I mean, our strategy is to grow, continue to drive innovation, and focus on building these long-term relationships with customers.
Jason Lopez: Rajiv Ramaswami is the CEO of Nutanix. This was recorded at the 2024 Next conference in Barcelona, Spain. This is the Tech Barometer podcast. I’m Jason Lopez. Thank you for listening. Tech Barometer is produced by The Forecast where you can find more podcasts, articles, and videos about the technology story and the people in tech. It’s at the forecast by nutanix.com. That’s The Forecast by Nutanix, all one word. Dot-com.
In this Tech Barometer podcast segment, McDowell shares insights from his report Taming the AI-enabled Edge with HCI-based Cloud Architectures and explores the impact of extending IT resources to the edge and the driving force of AI.
Find more enterprise cloud news, features stories and profiles at The Forecast.
Transcription:
Steve McDowell: The reason we push AI to the edge is because that’s where the data is, you know, we want to do the processing close to where the data is so that we don’t have latency. And in a lot of environments, if we’re ever disconnected, it’s going to shut down my business.
Jason Lopez: The question is, how do you deploy edge resources in real time? In this Tech Barometer podcast, Steve McDowell, Chief analyst at NAND research talks about his paper “Taming the AI-Enabled Edge with HCI-Based Cloud Architectures.” I’m Jason Lopez. Our aim in the next several minutes is to discuss how AI impacts edge computing.
Steve McDowell: We’ve always defined edge as any resources that live outside the confines of your data center. And there’s some definitions that say the extension of data center resources to a location where there are no data center personnel. It’s remote.
Jason Lopez: But AI, of course, adds complexity. One example McDowell cites is automated train car switching. The sides of train cars have bar codes which are scanned, and a local stack of servers processes where the cars are and where they need to be.
Steve McDowell: I can do this in real time. I can partition my workloads so that, you know, computationally expensive stuff or maybe batch stuff can still live in the core. And I don’t have to do that at the edge all the time. So I can really fine tune what it is I’m deploying and managing.
[Related: Slew of Changes Drive VMware Customers to Consider Alternatives]
Jason Lopez: This is important when you consider that AI at the edge differs from traditional edge deployments primarily due to its need for greater computational power.
Steve McDowell: Once we start putting AI in, then suddenly we have to have the ability to process that AI, which often means the use of GPUs or other kinds of AI accelerators. Ten years ago, if we talked about edge, we’re talking largely about embedded systems or compute systems that we treat as embedded. Embedded is a special word in IT. It means it’s fairly locked down. It doesn’t get updated very often. When we look at things like AI, on the other hand, that’s a very living workflow. If I’m doing image processing for manufacturing, for example, for quality assurance, I want to update those models continuously to make sure I’ve got the latest and the greatest.
Jason Lopez: And along with managing fleets of hardware and software in AI deployments at the edge, there’s also the issue of security.
Steve McDowell: By treating edge systems as connected and part of my infrastructure, and not as we historically have treating them as kind of embedded systems, if you will, it also allows me to, in real time, manage patches, look at vulnerabilities, surface alerts back up to my security operations center, my SOC. It makes the edge look like it’s part of my data center.
Jason Lopez: Tools like Nutanix allow for this approach, applying a consistent management practice across both core and edge environments. This involves deciding what tasks to perform at the edge versus the core due to constraints like cost, security, and physical space.
Steve McDowell: A key part of the conversation becomes what lives where? And that’s not a tool problem, right? That’s kind of a system architecture problem. But once you start partitioning your workloads and say, this certain kind of AI really needs to be done in the core, Nutanix gives me that ability and cloud native technologies give me that ability to say, well, I’ll just put this kind of inference in the cloud and I’ll keep this part local.
[Related: Pivot Past the Enterprise AI and Cloud Native Hype]
Jason Lopez: McDowell’s thinking springs from the flexibility afforded by hyper-converged infrastructure. The idea of AI at the edge is part of the whole architecture of storage, network and compute.
Steve McDowell: That can be as disaggregated as it needs to be. So if I need a whole lot of compute in the cloud, I can do that and then put the little bit at the edge and I can manage all of that through that single pane of glass, very, very powerful.
Jason Lopez: Treating edge computing as a part of the data center becomes so interesting because of how the data center itself is being transformed by AI and machine learning.
Steve McDowell: Once we abstract the workload away from the hardware, I’ve broken a key dependency. I don’t have to physically touch a machine to manage it, to update it, to do whatever.
Jason Lopez: The point McDowell makes is how management, not just of the configuration of a node, but across a fleet, is simplified. It enhances efficiency and scalability.
Steve McDowell: We’re taking technology that evolved to solve problems in cloud, but they apply equally to the edge, I think. It turns out, it’s a fantastic way to manage edge.
[Related: More Reasons for HCI at the Edge]
Jason Lopez: AI at the edge is increasingly adopting cloud-native technologies like virtualization and containers. The shift is to container-based deployments for AI models, sharing GPUs and managing them remotely.
Steve McDowell: If you look at how, you know, NVIDIA, for example, suggests pushing out models and managing workloads on GPUs, it’s very container-driven.
Jason Lopez: And McDowell explains why this simplifies edge management.
Steve McDowell: A GPU in a training environment is a very expensive piece of hardware. And giving users bare metal access to that, you know, requires managing that as a separate box. Using Cloud-native technologies, I can now share that GPU among multiple users, very, very simply. That same flexibility now allows me to manage GPUs at the edge with the level of abstraction that works. So I can sit in my data center, push a button and manage that box without actually worrying about what that box looks like necessarily. So I don’t need that expertise kind of onsite, right? Which is a key enabler for edge. If you have to have trained IT specialists wherever you’re deploying, that doesn’t scale. And edge is all about scalability.
[Related: The Future of AI Computing Resides at the Edge]
Jason Lopez: GPUs are typically what power AI, but are are not commonly found at the edge. But inference is a facet of AI that many technologists see value in at the edge. GPUs would be the right fit if at the edge, generative AI is needed. But what’s needed now are inference engines, especially around vision and natural language processing.
Steve McDowell: Take, for example, a retail environment where they have intelligent cameras that are positioned all up and down the aisles of the grocery store. And the only job that these cameras have is to monitor the inventory on the shelf across from the camera. And when they’ve sold out of Chex mix and there’s a gap there, it sends an alert, come restock. I mean, it’s very kind of data intensive and you don’t want to send that to the cloud necessarily.
Jason Lopez: Technology is moving toward managing infrastructure environments seamlessly, such as edge, data centers, and cloud, without changing tools or management models.
Steve McDowell: Nutanix has capabilities for managing AI in your workflow, kind of period, full stop. A good example of this is GPT in a box. Where it’s a technology stack and I plug a GPU in and I can do natural language processing. If I want to push that out to the edge. I don’t have to change my tools. I mean, the beautiful thing, and the reason that we use tools like Nutanix is that it gives me kind of a consistent control plane across my infrastructure. Now, infrastructure used to mean data center, and then it meant data center and cloud. And now with edge, it means data center and cloud and edge. The power of Nutanix though, is it allows me to extend outside of my traditional kind of infrastructure into the edge without changing my management models. So, as AI goes to the edge, I think the things that already make Nutanix great for AI in the data center are equally applicable at the edge.
Jason Lopez: Steve McDowell is founder and chief analyst at NAND research. This is the Tech Barometer podcast, I’m Jason Lopez. Tech Barometer is a production of The Forecast, where you can find more articles, podcasts and video on tech and the people behind the innovations. It’s at theforecastbynutanix dot com.
From the publisher's feed