All right! All right. How y'all doing? All right, I like it. I like the energy already. I'm pumped. But, and I love that saying there. I mean, simplicity is a superpower. That's, that's strong. And some of you might have heard this from me a long, long time ago before I joined VMware. I used to tell people that complexity is great for profits, just not theirs. And simplicity, it certainly can be the antidote, certainly a powerful antidote. And I've been thinking back. I'm starting to get up there, I guess. I've been working with VMware products and technologies now, coming up on twenty-five years, and I've never been more excited about our future than I am right now. You're gonna hear why during this keynote. When we're looking at the innovations that we're driving, it's all based on your feedback.
It's based on really listening and building our products the way that you've been asking us to do them for years. And that's what you're really gonna see today, is some transformational work that's been happening across the company. And when I look at this, I see so many familiar faces. I've talked to so many people already this week. This really is like a family reunion. It's just a different family reunion, right? You're not at the community center, you're not at your Aunt Janice's house. You're here, you're in Las Vegas, and that's a great place to be. And I think you know what? Your Aunt Janice, she'd probably wanna be here, too. So where are we at? You know, first, you already heard that quick introduction with VMUG.
What you may not know about our extended family here, our VMUG family is a 150,000 members strong in more than 50 countries. How about that? All right. 14 years as an independent organization and still going strong. Now, look, we have a lot to share today. You're gonna hear from some of your peers and how they're driving innovation. We're gonna deep dive into Cloud Foundation and all of the services that we're bringing along and around it. But first, what I want to start with is a perspective from our CEO, Hock Tan, and then I'm gonna be back later to talk about all things Private AI. Let's welcome to the stage President and CEO, Hock Tan.
Welcome. Welcome to VMware Explore. You know, it's great to be here. I'm amazed at the energy in this event, and it's great. It's been almost a year since Broadcom acquired VMware, and from conversations with all of you, okay, we've been hearing lots. And here's what we've been hearing from you: You're telling us you want our products to work better and be more user-friendly. You want them to actually work together. You're asking us, you're asking me, particularly: Roll up your sleeves, do the hard work, and you'll see over the next hour, that's exactly what we have done. Now, you're gonna see a change here at Explore this year. Sorry. Now, because we are serious business people, so are you. We are all about business at Broadcom, and we're here to help you run your business more effectively.
We're not here to show you bright, shiny objects, and talking of bright, shiny objects, 10 years ago, your CEO, your board of directors, fell in love with the promise of public cloud, and they drove you to public cloud first. Because of this, I see it, you're all now suffering from PTSD. You're confronting the three Cs of public cloud. Cost. Public cloud is much more expensive, more so than you can ever expect. Complexity. It's another platform. It's another extra layer for you to manage, and in compliance, if you have regulatory policy requirements, it's more complex, it's more expensive. In fact, a recent survey done over many global CIOs by a reputable global bank confirms more than eight out of 10 CIOs today are moving workloads back on-prem.
That's a huge change since the beginning of the pandemic four years ago. So here's my view, very simple: The future of the enterprise, your enterprises, is private. Private cloud, Private AI, fueled by private, your own private data. It's about staying on-prem and in control. Of course, you continue using public cloud for elastic demand and bursting workloads. But in this hybrid world, the private cloud is now the platform to drive your business and your innovation, and we have work to do to make that happen. Just look behind the curtains in your environment, which is why, in the first place, you moved to public cloud. There's plenty of room for improvement. You inherited legacy of data centers, which has created best of breed in compute, in storage, in networking. You're very siloed.
You are so screwed, because silos don't work well together, and it's painful for you to deliver services to your internal customers. When something breaks, as it often does, everybody is pointing fingers at each other, right? Makes you less resilient, days to find root causes. You want to deploy a new application, you need to write a ticket to your IT department, and you might just get that virtual machine two months later. In VMware, we believe we have the solution. We're taking the software we have grown to trust and love, ESX, vSphere, vSAN, NSX, and vRealize. We have made it all work together very well, and we will enable you, and we have enabled you in so many cases, to deploy it as a full stack to virtualize your entire data center.
In fact, to create this, a single platform, and we call it VMware Cloud Foundation or VCF. It's resilient, it's secure, and it costs much less than public cloud. We're not stopping there either. It's not just enough to build this nice Taj Mahal. We need furniture, and you want a rich catalog of microservices, like you do get in public cloud, to develop and deliver services at high velocity, and that's exactly what we've done. You see here, we have security, we have disaster recovery, automation, ransomware, orchestration, and now Private AI, a product just released three months ago. All these rich catalog of services run on top of VCF. You are, in fact, if you think about it, getting AWS on-prem. So before I close, I do want to come back to the three commitments we made when we acquired VMware. First, we're simplifying VMware.
We're simplifying the business, the products. From 8,000 SKUs before acquisition, we're down to four core offerings. That's all you need. Second, we're investing a lot to make these products easy to use and easy to work together, and to give you that robust catalog of advanced services to help deliver your applications. And finally, very importantly, we are investing to enhance the ecosystem for our partners and for you, our customers. Let me close with this: Your success represents our success. All of you here, clients, partners, friends, you are part of a community. You are the experts in private cloud, and as the pendulum swings back, your expertise in your enterprise, it will become more valuable than ever. With that, to dive deeper into VCF, it's my pleasure to introduce Paul Turner. Thank you.
Hey, great to be here, and what an audience! So you heard Hock talk about silos, data center silos, and how they're more expensive. They slow us down. They cost us. But let's really understand that a little more. The challenge is that it's not just about the technology. It's multiple security processes and policies. It's multiple approval processes and workflows, multiple upgrade and life cycle processes, complicated root cause analysis, poor capacity optimization, and much, much more. For leading businesses today, and that's all of you, IT is meant to be a business enabler, essential for success. But to deliver on that promise, IT itself must be agile. We have to be able to deliver services and applications quicker to meet business need. Let's break the silos of IT and enable an agile IT.
So IDC studied the impact of silos on cost, and you can see from their summary that it led to 42% higher cost of operations. That's people in this room. That is your cost of your people, that is your operations, that's your people cost. And it led to 34% higher infrastructure cost. That is capital expenditures, that's dollars that you spend on equipment. Now, it was not just you. We built products that were focused on each of the silos. We needed to do better. We had to bridge our silos. We didn't have single sign-on authentication across our products. We didn't have common policies. You talked about us not having unified tagging. You talked about us having independent product life cycle, independent logging, no common alerting. We could do better, and we have, and it is VMware Cloud Foundation. We unified all of our teams together.
Huge undertaking. We focused them on delivering the best private cloud platform for you, our customers. And VCF is more than just a technology upgrade. VCF is a strategic choice for you to deliver a cloud experience to your business, an agile cloud experience, but with the privacy controls, the security, and the resilience that you need from a platform for your applications. We have overcome our silos so that you can overcome yours. So VCF integrates compute, storage, networking, automation, and operations all into one unified platform. It's a software-defined platform for agility, but you don't have to take second best. Why? Because we're the number one market leader in software-defined compute with vSphere. We're the number one market leader in software-defined networking with NSX.
We're the number one market leader in hybrid cloud management with our management products, and we're a leader in hyperconverged infrastructure with vSAN. No second best here, and what's amazing is that if you do the TCO analysis versus the public cloud, you get 40% better total cost of ownership for running a VCF-based private cloud than a native public cloud, and that's based on our customer analysis over the past three years, so let's dive into it a little. You can think of VCF as really providing three layers: automation and operations for all of you to run it at scale, a modern, agile developer interface, and advanced services that we use to extend on the platform.
So the infrastructure and operations helps you, as virtual admins or cloud admins, automate, manage capacity, optimize it, do green data center efficiency management, all of your security patches, your life cycle, all unified into our automation and operations. Modern developer interface is all about agility. How do I deliver applications quicker? A cloud experience through service catalogs. Infrastructure as code with native Kubernetes, so that your developers can have the agile platform they need. VMs and containers, native in the platform. CNCF-compliant Kubernetes, native in the platform. AI and ML-ready applications, native in the platform.
The last layer we're going to touch on a little later are our advanced services, because if you've got this best private cloud that you can put in place, there's a whole ton of extended services we can provide with advanced firewall, advanced load balancing, managed data services, AI that we talked about earlier with Private AI, but we'll talk more on that later. VCF is a platform that we've also been innovating on, even in the last, the most recent innovations, and I want to highlight just a few of them. vSAN Max, ability for us to disaggregate storage and compute so that you can grow them independently. Live patching for NSX and ESX. All of you as admins out there will love this. The ability for us to take those security patches, apply them onto an environment, and do it without maintenance mode. Wow!
Able to actually hot patch ESX, hot patch NSX. VCF import, the ability for us to take all of your existing environments. You don't have to migrate them. Bring them over into VCF and get all of the lifecycle management that you want from the platform, and you don't have to move your workloads. Not just CPU virtualization, but GPU virtualization and DPU virtualization, and much, much more. We've continued to innovate. So all of you also want flexibility of choice. VCF is a private cloud platform, but it's not a private cloud platform that is restricted and limited to your data center. You have the choice to deploy VCF, the private cloud platform, think of it as an operations platform, into your choice of deployment. You can deploy it into the hyperscalers: AWS, Google Cloud, Microsoft Azure.
You can deploy it with our CSP partners, our managed services partners, like IBM Cloud. We've more than 300 of these partners. You can deploy it with our OEMs, who are building engineered solutions that are optimized for running VCF. You get the best experience with these optimized solutions, delivered from Dell, HPE, Fujitsu, Lenovo, Hitachi, and NEC, and more to come. And all of it is available through our vast network of distributors and resellers, with more than 14,000 partners who are there to help you. Not just help you buy the product, but help you implement the product and adopt the product and use it. So all of this is available today, but I thought, why not hear from a customer? So let's actually hear from GCI.
At GCI, the Enterprise Cloud Platform Group is responsible for providing the infrastructure and automation necessary for both our internal customers and our external customers to meet their business needs. Alaska is a very large state. We have a lot of unique challenges where we have to provide services to these different locations. Some of these locations can only be gotten to by airplane or boat, and those require unique engineering and other skills to get deployed and maintained. GCI employs a large number of the VCF suite of products. We particularly utilize the Aria Suite for automation, deployment of workloads, Aria Operations for monitoring and alerting on our infrastructure. Everything GCI does with VCF provides a huge impact to our customers and to the team supporting it. We're able to upgrade faster, provide security patches faster. We're able to deploy applications quicker and more efficiently.
Previously, it would take us months to deploy something out. Now, it can take us weeks or days. GCI right now is very focused on delivering a really good developer experience, so we're very excited to be talking about Tanzu Application Platform and getting that up and running for our developers. We're also very focused on looking at opportunities for Private AI to provide better security and keep that data in-house. By removing a lot of the operational burden, VMware Cloud Foundation allows us to focus on our application developers and providing new functionality to our customers and increase business value.
Excellent, so I love the fact that you don't have to take second best in Alaska, and actually, Jeremy is here with his team, so Jeremy, I'd love you to stand up, and let's give a round of applause, truly, for Jeremy. It's something that all of you as customers can achieve, too: build and break down those silos, but we are just getting started, and I am delighted to announce today that we're releasing a new announcement for our VMware Cloud Foundation 9. VMware Cloud Foundation 9 is a unified platform to break down those technology silos. An agile, automated cloud experience for all of your workloads, and a secure platform, trusted for your most critical, confidential workloads, helping protect against malicious actors, disasters, ransomware.
So it's a platform for helping you build, deploy, and operate and consume. We've integrated into a VCF, a new VCF operations console, all of our events, logs, diagnostics, and we're allowing you to root cause analysis issues quickly and maintain continuous service uptime. This means your Network Insight logs integrated, your Log Insight logs integrated, your ops logs integrated, everything into one place so that you can root cause analysis and correlate issues. We extended this with Skyline. Skyline, we've actually taken and integrated it now back into the product. We take the rules that are coming from all of our common issues, we push the rules into your product, we analyze for those common issues, and then if we can't discover the issue, we auto-publish logs back to VMware for further analysis so that we do it all automatically for you. We centralize governance for policies.
OAuth brokered identity is a common identity system across all products. Unified license management, unified fleet management to allow you to plan, schedule, patch, and upgrade your entire private cloud, and we also are very focused on agility. Agility also means that you need to look at: how do I provide tenancy into the platform? Because if I can actually manage different tenant groups or different application groups, I can set business and service policy controls on those applications, then I can do chargeback and accounting based on what each of the application teams are doing and using, all integrated into a single new VCF automation console. Application blueprints to build out a service catalog with multi-step workflows and a very simple low-code interface for development.
And a native infrastructure as a service, which extends beyond the Kubernetes that we just talked about, the runtime Kubernetes, because we brought in credentials management service, a storage, networking, and load balancer service native in the platform, a registry service, a secret service, a data services management for persistent applications, and integrated monitoring, all now as part of one unified VCF platform. So we also extended the core. Advanced memory tiering. You look at the cost of systems today, memory is taking about 50% of the cost of the system. Imagine you could tier that very expensive and high-cost primary memory and tier it out to NVMe. Well, you can. We can dramatically reduce the cost of memory on systems. vSAN deduplication, vSAN ESA with multi-site replication and deep snapshots, and I'm really happy with our native VPCs.
These VPCs that we built out with NSX, we've actually integrated now into vCenter. It's an extension of vCenter's networking. No more management of NSX manager, no more management of edge servers. You can deploy VPCs natively. So for many of you, you're probably thinking: "Well, how can I get there? It seems daunting." But VCF allows you to get to where you want to be at your own pace, and the best way is to take a private cloud maturity model assessment. Scan this. Please join us on Wednesday. We have a session to actually go through it. This. And also, you'll be able to sign up for this. All of you should take this. Very quickly, an outside-in assessment of your needs, mapping to your business outcomes, and defining a pathway for you with the highest ROI.
So one thing for me to talk about it, but let's actually hear from a customer. CPS Energy serves 1.5 million customers in San Antonio, Texas. Please join me in welcoming CIO Evan O'Mahoney to the stage. Hey, Evan.
How are you [ audio distortion]
So thank you for joining us, of course, first. For starters, tell me about your role and your team's role at CPS Energy?
Yeah, absolutely. So CPS Energy is the largest, municipally-owned gas and electric utility in the U.S. We're a fully integrated utility, so we manage, own, and operate our generation assets and also control the transmission and distribution parts of the system as well, all the way to our end customers. As the Chief Information Officer for CPS Energy, I'm responsible for all aspects of our technology strategy, and I also sponsor our digital transformation program, which is branded Evolve, and that transformation for us is really focused on improving the experiences for both our employees as well as our customers.
Excellent to hear. Now, you're actually driving a multi-year cloud transformation, and it's centered on VMware and Azure. So tell me more about that.
Yeah, so, you know, about two and a half years ago or so, we were really trying to figure out how to build a flexible architecture to support the evolving needs of our business. You know, I mentioned our digital transformation strategy, in the early stages, had quite a bit of ambiguity in terms of what we were actually trying to accomplish, but what we knew is the underpinning for all of our technology strategies needed to offer flexibility and agility to help navigate the future for us, so as we looked through kind of our current state, we had quite a bit of legacy infrastructure sitting in our private data centers. We were already a VMware customer and leveraged that partnership to transition to VCF in concert with Azure VMware services as well.
So that gave us a lot of flexibility to deploy new infrastructure in our private data centers and extend that private cloud experience, but also begin migrating workloads where it made sense to Azure as well. And we were able to do all of that while continuing to maintain a single pane of glass for our engineering and technical teams.
Wow! It's very powerful. So you've got a private cloud platform, but you get the agility because it's a common cloud platform on Azure and in your data center, and it gives you the flexibility of choice. Now, we talk a lot about the technology, right? But people are a huge part of it, I know, for you. And of course, how did you work with your team and leverage their existing skills, and kind of how did they adapt to this?
Yeah, so our people are at the foundation of everything we do at CPS Energy, and one of our values is one team, and we really take that value and use it as a lens as we're making all decisions as an organization, and for me, that's certainly true from a technical perspective, so I really think about people in two different ways. First, we have our technologists and our engineering team, and I'm extremely lucky to have a group of such talented, you know, technical team members that really are able to effectively manage all aspects of our ecosystem.
As I mentioned, really being able to leverage the experience and capabilities that they had to accelerate our transition to new technology platforms and into the cloud, while also building out that longer-term strategy to really get them focused on how we can improve our automation and orchestration in the future, whether that is creating, you know, more automation in our private cloud world or leveraging the orchestration to move workloads back and forth between the public cloud and our private cloud. You know, making sure that our team is equipped with all the knowledge, skills, and capabilities. It was certainly a key focus for us.
But it was your existing team, and you're able to build based on their current skill set, and they were able to evolve and build out a full private cloud with you. Now, you've got thousands of these IT practitioners and admins, and cloud admins or virtual infrastructure admins in the audience here. What would you like to share with them?
You know, maybe two pieces of advice as I reflect on our transformation. The first is, don't underestimate the importance of building an effective plan. You know, for us, our environment had thousands of virtual servers supporting 500+ applications and countless numbers of integration across that landscape, so quite a bit of complexity. As we went to embark on this journey, really making sure that we were diligent in laying out a process to migrate to our future state in a way that would limit operational disruption. So really spending the right amount of time on building out that plan, critically important. The second thing that I will say is communicate the value to your customers, right? So we have 1.5 million customers that we support, but I also have an organization of internal customers that are consuming our technology services.
And you can do all of this great transformational work, but the technology isn't going to explain to your customers, internal or external, the value that they get. That's the responsibility of IT. So being able to translate all of the benefits, resiliency, automation, orchestration, speed to deliver, in a way that your business stakeholders understand, critically important, and I couldn't stress enough how that really contributed to our success.
Excellent, well, thank you so much, and actually great to catch up.
Thank you.
Thank you all. So I talked about a final layer. We talked about the platform, a confident, secure, private, resilient platform that you can run anywhere. But the final layer is our advanced services layer, and you heard Hock talk about that a little earlier. But let's have a little more detailed view of that. It's a robust catalog of services that are ready to deploy on top of the VCF foundation that you already have. And, you know, to bring that to life, why don't we walk through a day-in-the-life scenario that'll be familiar to many of you? And with that, I'm gonna hand over to Valentina and Chris, two of our senior technologists here at Broadcom.
That's great. We're excited about getting started. We look forward to talking with you two again on Thursday. Thanks, Lisa. We'll be in touch then. Bye. Bye-bye. We have all been there. We have a long list of requirements, and now it's time to map out a journey to get there. In this hypothetical scenario, we're going to think about ACME's Corporation's journey. Chris, I think we have a great understanding of what ACME wants to achieve. They are eager to get started with their consolidation of apps from the public cloud to the private, powered by VCF cloud. By the way, cool shoes!
Thanks. I was actually just about to, you know, say you've got a great coordination going on with the orange, but I couldn't pull off the orange and the pink, so I went with the VMware blue and green with the Broadcom red swoosh. You know, I call these my acquisition colorways.
That is totally awesome.
Yeah, but look, I don't wanna go down the rabbit hole of talking about sneakers. Let's get back to ACME. You, me, 5,000 of our favorite engineers here, let's talk about ACME and their pilot app. So we know that this pilot app is five microservices. It's got some data services, a database... Actually, look, you got me all excited about talking about shoes. You're the app person, so why don't you talk about the app?
Well, what you said was absolutely spot on, but what ACME is also trying to do with this app is that they want to make sure that the app is always on, and so they're designing it for full resiliency, so any failure at the infrastructure should not ever affect the app itself. They also made sure that they design it for auto-scaling, so as the end users' needs change over time, the app will adjust automatically. And last but not least, they have a global customer base, like all of you here, and so they wanna make sure that the app just sort of follows the needs of the users on a global scale. You know, they're a little bit iffy in this move from public to private. You know, they were sharing with us that they really wanna make sure that their requirements are preserved throughout this move.
Yeah, that's understandable, but it seems like ACME has a very mature software development practice.
Yeah, totally. I mean, they make use of pretty much every possible public cloud service, and that really drives their developer productivity. Just to give you a sense, they, they push code to prod once a week.
Wow, that's often.
Yeah, yeah. But hey, I'm confident. I think we can make them successful. So why don't we review the plan that we have in place, layer by layer of the architecture? Maybe we should start from the infrastructure, but you are the double VCDX guy, so Yeah yeah, of course why don't you start us there?
Yeah. Yeah, of course. You know, look, we know that ACME's already deployed VCF, so we know that they have a private cloud platform that will support traditional VMs, containers as they need. But we're not talking about just deploying them. We're talking about using VCF, which is purpose-built to simplify the provisioning of really all of the constructs that are needed. So we're talking compute, storage, networking, load balancing, security, all wrapped into workflows that make it easy to deploy, but better yet, disaster recovery being built underneath all of that.
Yeah, and they also started making use of the advanced services. So the infrastructure operations team has started using Tanzu, for example, to really drive their large-scale container operations and Kubernetes cluster management needs. But they also started taking advantage of the rich data services and developer services, so that the application team can easily build, deploy, and scale an application as they need it. And as a Tanzu product manager, I'm really excited to announce Tanzu Platform 10, which brings a rich set of capabilities for platform engineering teams to drive application delivery in an AI-ready, self-managed fashion.
Look, as excited as you are about Tanzu 10, I'm excited about the fact that ACME, not only are they evolving their technology stack, but they've really made significant changes to both their people and their processes.
Yeah, yeah.
They're heavily invested on building a dedicated cloud platform engineering team.
Absolutely. They did put a lot of thoughts into how to blend together the right set of skills. So initially, of course, they started from what they had in-house, which was a lot of strong infrastructure, networking, and security expertise, and then they blended in a strong SRE practice, and as of lately, they blended also in their DevSecOps practice. You and I spend a lot of time with customers. We know that that's key for success of building a strong private cloud engineering practice.
Yeah, very key.
Yeah. Speaking of people, I recently met with their new PM. She's awesome, but I think what she's really driving, it's a shift in mentality and thinking about platform as a product.
Yeah, she's been instrumental, really, from a top-down approach in getting everybody to buy into this idea that the lines of business are really customer zero in the private cloud.
Yeah, yeah, absolutely. The combination of VCF, advanced services, and their strong platform engineering practice means that their application teams are gonna have access to all the primitives and the agility that they expect from the private cloud. But at the same time, what is instrumental here is that the platform engineering team has a much higher level of control, what dev tooling are used, what processes, cost, data locality, and so much more. So I think we're set there.
Yeah, it seems like the next piece of the puzzle is to introduce App Spaces.
Yeah, absolutely. So App Spaces are a powerful way to dynamically provision an application runtime.
So hold on. When you say application runtime, that's just kind of a fancy way of saying that everything that application needs to function and run is provided by this App Space?
Yeah, yeah, well said, and what is neat about App Spaces is that different personas can contribute a little slice of that definition. So for our pilot app, the application owner said, like, "Hey, I need an API gateway, and I need a database service." The security persona contributed the fact that they wanted all traffic between any components of these applications to be encrypted, and the SRE team said, like, "Hey, whenever you deploy an app in production, in order to be resilient, it has to exist across two availability targets," and App Spaces just sort of magically make it happen, and so that's how we guarantee that their requirements today are preserved as they move to the private cloud. Now, the space is ready for the app.
But hold on. You didn't mention anything about the resiliency or the front-end security. How does that kind of play in?
Yeah, yeah, so that was sort of already baked into what I just described. They're making use of our advanced services for load balancing. So Avi contributes a strong load balancing capability, WAF, and container ingress. And the combination of Avi, VCF, built-in resiliency, and the Tanzu App Spaces makes sure that the app is resilient, it's available in a multi-region fashion, and auto scales. So app team is good. What about the security requirements?
Actually, you know, security's been really interesting with ACME because they've had a significant change in their mindset when it comes to security. They're not looking at it as security as part of one particular team, but how does every silo become a good corporate steward in the overall security posture? So they started by using vDefend firewall to build some infrastructure rules that mitigate against the most common things that an attacker wants to do. So we're talking about protecting against SSH, SMB, TLS, protecting Active Directory, DNS, or probably the number one thing that attackers want is RDP, or the Ransomware Distribution Protocol.
Yeah, that was a lot of acronyms.
Yeah, it was, but I'll actually have some more for you later.
All right.
Yeah, but look, in addition to the infrastructure rules that they built, they've also used a simplified tagging mechanism to be able to control the way traffic flows between different zones of their environment. ACME's really found out how to quickly realize value and buy down risk by controlling flows at a higher level, not really necessarily getting to micro-segmentation yet, but figuring out how do we get something in place that buys down risk. We can control the way things communicate between prod, test/dev, and DMZ really quickly and easily.
That's, that's neat. I know the application operations team was really keen on having some of this strong separation between development, staging, and production environments, so that's great to see already there.
Yeah, but you know what? That's not all. I feel like I'm selling something on QVC. That's not all. When it comes to your app, we can actually use security intelligence to monitor your app, right? So we'll watch it for a period of time, and it'll give us back all of the flows from the different components of your app, but it'll also provide us recommendations. Now, this is not AI ML. This is a human looks at it and says, "Okay, these all look good," but then we can easily convert those recommendations into rules and give you a fully micro-segmented version of your app.
I feel that's a much higher level of security than what they get today on their private cloud.
Oh, I agree completely. But look, I promised you more acronyms, so here are the ones that CISOs and their crew love: IDS, IPS, NTA, NDR.
All right, folks, it's official. Chris wins the gold medal for acronyms.
Thank you very much. Thank you very much.
Hey, you threw a whole lot of security in there. Is there any performance impact we need to worry about?
You know, it's a great question. It's actually a very common question, but the answer to that is really you have to look at the architecture. And if we were leveraging some architecture that was hairpinning traffic to some physical device, I would say there's probably a good chance we're gonna have an impact. But with VCF, you know, we're leveraging the distributed architecture. We're building security into every node in the data center. So as you scale the data center, you're not just adding more CPU and memory, you're adding more firewall capability. You're adding more advanced security capabilities with every node, so the performance impact would be negligible.
Yeah, that's the beauty of distributed systems, I guess.
Yeah, but you know what? There's more. ACME is looking to really take security to another level, so they've introduced IDS, IPS to protect against known bad signatures, and now they're starting to use network traffic analysis, network detection and response, and now we introduce AI and ML. So we're looking for anomalies, the different ways of communicating from a baseline, or mapping attack campaigns for known attack methods, blocking malicious downloads, or maybe most importantly, is giving the security operations team all of the data they need to be proactive in the event that there is a cyberattack.
Fantastic. I think the CISO and his team are gonna be really happy with this. This was a fantastic run-through, and I'm confident to say that we meet the needs of the application and the team, and the app can now move.
App can move.
Woo-hoo!
All right.
That's awesome.
Yeah. Well, actually, let's not get too excited yet. We know that there is an additional requirement-
Oh, right
... to scale this app to what? 250 ish retail locations.
Yeah, that's right. So why don't we get someone from the Edge team to help us think through the requirements for that?
Edge team, Edge team. Sanjay.
Okay, let's go.
Let's see, Sanjay.
Thank you, Valentina and Chris. So now let's look at extending this architecture that we've seen to the 250 retail locations that ACME has. These are the locations at the front line of the business. This is where ACME's business gets conducted. So the first thing that we're going to do is we're going to extend out that architecture across the network using software. We're gonna define a software-defined wide area networking technology using VeloCloud. And what VeloCloud is going to allow ACME to do is get the rock-solid network resiliency and provide the best user experience, both for those personnel who are in the store, as well as for the shoppers who walk into the store. And that's all going to be done with simplicity and edge-based orchestration. So edge-based orchestration is gonna satisfy the needs of the IT community.
Now, you can think that across these 250 globally connected edge locations, there may not be some locations where a wired infrastructure is available, but VeloCloud can support that same network resiliency regardless of the underlay. So using fixed wireless access on 5G or the recently available satellite connectivity, VeloCloud can optimize the traffic on those kinds of underlays as well. And the best part of this is you've heard a lot about AI. VeloCloud is AI-ready because the nature of traffic that AI brings is very different from the traffic that we've seen today. With almost 500,000 endpoints that we've already deployed and 5 trillion data points that we see every year, we can recognize the details of AI traffic, and because it looks very different from today's traffic, there's a lot more data being generated from every one of those edges.
That data is multimodal, and so all of that data going upstream, that data has to be optimized for low latency so that ACME can continue to get to the business goals that it has. Now, let's look into one of those stores. What's happening inside the store? So what ACME wants to do is that it wants to deploy workloads in the store itself. That is both for the current applications that are available, like point of service and inventory, but also for connecting these smart devices that you see there: video inferencing cameras, Zebra scanners, self-checkout kiosks. All of these produce data, and they have to be handled. The best way to do that is to roll out an edge compute stack, and the edge compute stack is purpose-built.
That purpose-built edge compute stack, what it does is it allows ACME to have the workloads that are there for today's applications but also prepare ACME for the Edge AI workloads of the future. What Edge AI brings is, again, a different traffic pattern, but now you can think of a shopper that walks into one of ACMEs stores. That shopper is greeted by a supervisory shopping agent to make her shopping experience the best. That supervisory agent can talk to an SLM or a small language model that runs on that edge compute stack, and that agent talks with equal ease to the SLM as it does to the LLMs that are located back on VCF using Private AI. So there you have it. We have that architecture that you've heard so eloquently from Valentina and Chris.
That architecture has now been extended to each one of these 250 retail locations that ACME has, much to the delight of the shopper that walks in, also to the satisfaction of the IT personnel, and the line of business gets increased revenue. Now, we've shown you this entire catalog of advanced services that run with VCF. Those services, networking, security, compute, bring all of this together for ACME. But let's look at a real-life example, the U.S. Senate Federal Credit Union. We're gonna walk through how this works for them. With this, let's watch the video.
At the U.S. Senate Federal Credit Union, our customers are our members. They have direct ownership in our organization, and they've got very real interest in our success and what we can innovate on. VMware Cloud Foundation is a central part of our team's strategy that brings together all the different aspects of running our own private cloud environment and enables us to operate such an environment at a scale and sophistication that we otherwise wouldn't be able to achieve.
VMware Cloud Foundation is kind of those Swiss knives. You have plenty of different tools, and they're at hand when you need them.
In a private cloud environment, we can provide a level of security that we feel is unmatched with any public cloud provider.
In this organization, we are exploring a new world where we can build our application, where we can use AI to create more technology, where we can use machine learning to provide better services.
What we've really looked for in Broadcom is a partnership that allows us to explore any opportunity that arises quickly and effectively.
Private AI is very important for us because it allows us to play with our own application.
We see AI and customer service as such a powerful use case, everything from knowledge-based powered chatbot answers to our internal staff, all the way to delivering financial answers to our members on whatever device they use for banking. VMware Cloud Foundation was instrumental in meeting our business objectives to begin evaluating artificial intelligence within our enterprise.
When it comes to VMware Cloud Foundation, I can say that we were able to create that in a couple of days. Then the part of Private AI, I would say another two days, maybe three. Everybody who is listening may be, you know, saying, "Oh, if that guy could do it, why can't I?
Please welcome back, Chris Wolf.
All right. Yeah, I really love the ending of that video. Like, why can't I? And that's really what we wanna talk about here today is this really is attainable for everyone out there today, and what you just saw with U.S. Senate Federal Credit Union is how quickly we were able to work with them and deliver value. Now, last year, you might remember, we introduced Private AI as an important market segment, where we were looking to balance the gains that you get from AI with the privacy and compliance needs of the organization. What we've seen since then is we've launched a product just past May. We've seen tremendous uptake, not just in financial services, but in a lot of industry verticals. Public sector, we've seen it in oil and gas, we've seen it in retail.
Lots of industries are really working to partner with us to gain these benefits. And the bottom line with Private AI, if you just kinda simplify things, we're taking the AI model, we're taking the AI service, and allowing it to run anywhere you do business. So we're bringing the AI model to your data, keeping it safe and secure. Now, let's get down to this, though, right? You're like: "Well, that sounds great, Chris, but it's not that easy." And you're right, it's not. Developers and your data scientists, you know, they're A/B testing models. They're looking to integrate models with their applications, making sure that they have the resources they need so their applications are performant and scalable. Then on the operation side, it's like, you know, it can be terrifying. "Hey, you know, I need to make sure that these models are safe and secure.
I want a way to govern the models and make sure that they're approved for production before I'm introducing these things. I need to maintain availability, scalability, right? Security and compliance. So there's lots of challenges here. And then the other thing you might be asking is, "Well, why VMware Private AI? Like, there's a ton of solutions out there. Why us?" You know, first and foremost, it is privacy. I wanna gain the benefits of AI without losing control of my data. Speed and agility. This was one of the ahas we had. We are so good at automation that sometimes we take it for granted. And in this case, what we realized at the NVIDIA GTC conference earlier this year is people loved the fact that we can take these complex AI applications, dozens of microservices, and be able to stand them up in minutes.
We've heard from our customers, the process that they are often doing manually takes them sometimes weeks to months to get these applications online. Now, something else you might not be aware of is. Well, let me back up. Actually, you're our base. You are absolutely aware of this, but you may not think about the correlation to AI. When you think about AI, I'd say the other aha moment we've had with our platform is that customers that have been deploying these resources on bare metal for years, in some cases, have come to us and are ready to deploy this platform today. The reason why is, as you all know, distributed resource management at scale is extremely complex. It's really hard to do. And guess what? We've had DRS, our Distributed Resource Scheduler, now for almost 20 years. It is hardened and proven technology.
It can help you to load balance and automate and scale all of this complexity we have with AI, from sharing GPUs, to networks, to memory, to making sure that my performance mandates are met. So really powerful technology and a key differentiator we have. Now, you get the centralized operations. You have the same set of tools and processes for AI and your non-AI applications. And in all of this, what's really empowering as well, I think, is you get a lower total cost of ownership. We have had customers tell us that running AI services on our stack can be up to one-third the cost of comparable public cloud AI services. So you get a ton of benefit here, and you also get the lower cost. Now, you might say, "Well, that's great, but come on, there's got to be a catch.
If I'm gonna virtualize, then I'm losing performance, right? I got to be sacrificing performance here." Well, check this out. So what you're seeing here, first, the baseline is bare metal performance, which is one. So what you're seeing here when we're virtualizing GPUs, and there's many cases here in this benchmark, we're exceeding bare metal performance. Now, this is using MLPerf. It is the standard AI benchmark that's in the Linux Foundation. It's what everyone is using, and this is 8-way H100. 8 H100s on one server. In many cases, we're exceeding bare metal performance. Why is that? It's because of our scheduler. We have the most efficient scheduler in the industry. We, for decades, have had to run multiple operating systems and applications on the same server continuously, and we have the best in the industry, and it shows right here in this benchmark.
Now, this is a solution we introduced last year. Many of you are already adopting this, and this is VMware Private AI Foundation with NVIDIA. It combines the very best of VMware technologies and the very best of NVIDIA technologies to deliver a turnkey AI solution to bring you very quick time to value. There's just lots of goodness here. You see a lot of choice for OEMs around it, and the key here, too, that I didn't mention, is this is a platform, and this is the way you want to bring AI into your environments. You take a platform approach. Why? Because the industry is moving so fast. The AI model that was great for an application last week may not be the best choice tomorrow or next week. Taking that platform approach allows you to iterate via software.
As new things come along, I can quickly onboard them. I'm not locking myself into a particular service stack. So again, tons of benefit here, and again, above the stack, I can do things like run NVIDIA inference microservices really quickly and efficiently. So that's just natively integrated, run them in production. I can bring other alternatives, like our own, like Spring AI, on board or any other third-party services that you're thinking about as well. We continue to grow our ecosystem. We've announced more partnerships this week as well, companies like WWT and HCL for expanded services to help you really get value out of your AI applications and do so quickly. New code assist partners as well. These are companies like Codeium and Tabnine, continuing to grow that ecosystem where you guide us. Now, let's talk a little bit about some new innovations.
So first up is the Model Store. I mentioned the importance earlier of being able to govern and control models. Now, the approach we took here is really important. You didn't wanna have another tool to do model governance. Harbor, as a container registry, is the right tool and the right approach to do this, and you're starting to see others do this as well. So this is how I can quickly integrate models from Hugging Face using the Hugging Face CLI. I can integrate with NVIDIA NGC. Any custom model you want, this is how you're gonna be able to bring these into production safely and securely, share them, and collaborate as well.
There's a lot of additional future capabilities as well that we're rolling out, and this, this includes automation for how we're standing up the stack, lots of work around GPU management and reservations, and also other core AI services that I'm gonna show you here. So this is cluster-wide GPU visibility, right? GPUs, they could be hard to come by. We wanna make sure we're maximizing every ounce of compute capacity of our GPUs that we have, and this is how we can keep track of them in real time. Now, the other thing we've introduced now, we're introducing now, is our Data Indexing and Retrieval service. So all the rage right now, you're seeing lots of folks building these RAG applications, Retrieval Augmented Generation.
It's allowing me to take a foundation model, connect it to my most recent data sources, and be able to populate that into the model so that I'm adding new context, I'm improving accuracy. I'm able to do so with very little investment on my end. This is gonna allow you to take your models, connect them to your different data sources, and be able to populate all that. We also even have a refresh policy, so that you're able to continually re-index at the time that you feel is appropriate for your organization. So now we've laid out models, being able to deploy 'em, govern 'em, connect to data sources. The last part, which is really exciting, is our Agent Builder.
Through the Agent Builder, you're saying, "Well, hey, how can I now build a chatbot so that now I can start to interact with that as well?" This is what you're gonna be able to do with the Agent Builder. Just using natural language, I can now communicate, I can build that chat interface, I can start to test it, I can roll it out. It really is that simple, and you just heard that from US Senate Federal Credit Union. We're really driving this with our partners to really make this simple and consumable for you. Oh, there's one more thing. GPU high availability. Sounds easy on the surface, but nobody actually wants to have GPUs sitting idle for high availability purposes, right? They wanna be able to maximize them all the time. What you're seeing here in the vSphere Client is I had a failure.
Server's out. I need to restart this production inference application somewhere else. So here's how we're doing it. So you see through VCF automation, that's restarted. It's going well. So where did the capacity come from? So what we did here is, through using VM classes, I could do things like have our data scientists experiment or research workstations at a lower class. I can gracefully shut them down. I can now use that capacity for things that I need, like my production applications. So we're taking a new, fresh look at HA, super exciting, really happy to share that with you today. Last but not least is Ethernet for Private AI networks. You know, this is something that Broadcom obviously has been driving since its entirety here. And why does this matter? This isn't even the future. This is today.
Six of the top seven hyperscalers, their internet or their network backbone for AI is entirely Ethernet-based. Ethernet can recover 30% faster. It has the broadest ecosystem in the world, right? When you think about network topologies right now. So there's tons of benefit, and I encourage you to talk with your Broadcom team members that are here at the conference and/or your account teams back home, about how this can help you to move forward at a far lower cost. Okay, that was a lot of tech, right? We covered a lot here today. So let's just kinda talk about, you know, what all did we see. VCF 9. VCF 9 is a transformational platform. We reorganized ourselves internally to give you a single unified product, one product that's gonna allow you to run a public cloud experience anywhere you do business.
We introduced Tanzu Platform 10, and how easy it is to start to build and integrate AI into your applications. Our Private AI Foundation with NVIDIA is what I just shared here. Tons of momentum. We have lots more sessions that are gonna cover that, as well as all of these topics today, and software-defined edge innovations, how we can use this now to make it so simple to manage, operate our AI applications at our edge sites. Last thing, now, you've taken a lot of pictures in Vegas. You've taken some interesting ones. This is the most interesting picture you wanna take right here, and you might ask why. This is where you can dive deeper. Really go deep into VCF, go deep into Private AI, go deep into Tanzu and what we're doing with advanced network security. Check out these sessions.
We've opened up some exits here at the back. Please exit this way. Have a great conference. Hope to see you around.