STATE OF SYSTEMS Ep3: BROADCOM - Addicted to Cloud - the Broadcom Intervention with Joe Baguley

When the Broadcom acquisition was announced there was a broad range of reactions, shall we say, at the marketplace. The M.W. itself was an absolute hotbed of innovation. We literally were, you know, patents popping out every other day and we were kind of more like a religious leader. You know, the new gadgets, the new features, the new bits that we're doing. That's so exciting to me. We reorganized the whole of R&D, flipped it on its head. It used to be sort of like a parts company and now you're actually selling the car. Yeah, completely. We went from 42 to 7 data centers in 14 months. And we went from 2.5 thousand people managing up a private cloud to 68. So how does VCS9 help? That's how it helps. Cloud was supposed to save us, swipe a card, spin up the servers, job done. It was meant to make IT simple, cheap and limitless. But 20 years later, the bills are racking up. Compliance is an absolute nightmare. So ask yourself this. Do you even know where your data has been stored at the minute? Most can't answer that. Now Broadcom's made its move. 61 billion dollars. VMware, cut down to one product. Customers suing Fortune 500s, signing decade-long deals. So what's the truth here? Is Broadcom breaking VMware or forcing the future the rest of us were too scared to build? So today we're asking why hasn't public cloud delivered? What does Broadcom's big VMware reset really mean? And what should IT leaders do next? It's got through this. We've got Joe Bagley, a mere CTO at Broadcom. For more than a decade, he's been in the room with banks, governments and global giants. Helping them figure out where their workloads really belong, public, private or both. Joe, welcome to state of systems. It's good to be here. Before we get started, we have a Hacker handle. So we ask all our guests, which is your first pet and your latest illness or ailment. Mine, for example, is Andy Galt. [LAUGHTER] Andy Galt. All right. Well, mine would be Rupert, and I can't remember my last-- I don't know, really. ailment. You just coughed a minute ago. I'm deaf and should be wearing a new-- Rupert Flu. Rupert, Rupert, deaf. Hello. And the mine would be Ben Shoulderpop still. I think I need to go to the physio for that one. It's been three episodes now, and I still. My shoulder is still really bad. But yeah. Joe, can we take a bit of a walk back? Everyone rushed into public cloud, offered without really thinking about it. Some workloads thrived. Others were a disaster. Can we go through the classics with you? Big ERP systems got lifted into the cloud in the early rush. What actually happened when people tried that? Well, look, it's just-- I'll say it now, and it will come up again. But every conversation about technology ends up being one about people in process. And ERP systems always been the biggest chunkiest thing in most large enterprises. It's what they spend the most. They seem to think it's the most important application. Actually, to be fair, the most important application in most open enterprises is email and spreadsheets. But in reality, ERP is what they think they care about. And so they've always spent big money on that, and moved it across from mainframes across to the mini computers. If you follow back through history, ERP systems, it ran on the mainframe. And then mini computers came through. AS400. So we move it all to AS400. And then all unix is around. So let's move it all around it on our 6000, if we're going through the IBM product. And then, oh, let's move it onto Windows, because everyone's running stuff on Windows. Now, now the cloud's along. So let's just move it onto cloud. Because that's what-- And you kind of get this idea that that's almost what IT got in the process of doing was just re-platforming stuff. And so when cloud comes along as the next platform, it's almost like, well, surely that's what we must be doing now. We're always putting this onto the next platform, which is cloud. And equally, I think CIOs always look for that simplicity. They're always looking for a way to make it much easier and simple to manage stuff. So if they go, if I take it and I move it all over there, then I can shout at someone else about it almost. So it's going to be simple. It's going to be easier. It's going to be less hassle for me to manage than the stuff I'm doing on-prem. And of course, when we came along, every platform promises you it's going to be cheaper, faster, all these kind of things. And so of course, people get taken in with it. And every time, every single time, what we end up with is we realize they're not always cheaper and faster. What it is, it's a different platform for a particular type of workload. And so you end up with this decision of the right workload, right place. And I talked to a large enterprise this today that still have mainframes. Right? Got a huge mainframe business. And they still have huge window service dates and still have Linux and Unix running everything that they do and have stuff in the public cloud. So it's this, shall we move everything into one place? It's never going to work for anyone. Yeah. And is inertia a big thing to those large corporations? I think inertia, but inertia is all about organizational inertia. Right? It's about the fact that there's people that slow things down. That's actually what happens in most organizations. So it's rarely technology or technology that holds people back. It's usually people that hold back what we're doing with technology. It's, I've got this skill set. I've got this organization, you know, Conway's law starts to fire up all the time. You know, then their comms channels start to mirror what they're building and all that other kind of thing. And it's almost like you see this horrendous mess time and time again. And you know, one thing that blew me away, which we can talk about later actually is what Broadcom ourselves did following the acquisition of VMware and reinventing RIT because it's one hell of a story that you wouldn't. Most people don't believe until our CIO sits there and tells you it personally. It's quite shocking. So I mean, Segway's into AI perfectly. Obviously, we're talking about people and now we've got AI coming through. Where's the benefit to the companies? Because obviously everybody talks about it. But actually, is it just a quick chat GPT lookup or how does that actually come into the large corporations? Again, so you see these cycles, right? So the last big hype cycle was blockchain and everything blockchain. Everyone's got a blockchain. I'm going to put everything on the blockchain. I'm going to solve everything with blockchain and everyone is just another database. Right. And you go, well, could the blockchain, how should you speak, could you use a database? Yes or no. If yes, then don't use blockchain. It's almost the same with AI. It's this next big hype cycle of AI is going to solve everything. AI is going to be brilliant. It's going to do, you know, if I'm not in AI, I'm not cool. And that's almost the thing. It's always the CIO or the tech lead or even the CIO. I was concerned if they're not talking out to their shareholders going, "Yeah, we're investing in AI because they're going to look bad and it's not going to work for them." So they always have to be seen to be doing it. I think like any technology, we then very quickly start to realize the limitations of that technology and start to realize where that technology should be applied and where that technology shouldn't be applied. And I think, you know, when you look at blockchain, we took a while to get to that. There was a fair amount of hype and we went through an awful lot of discussions around that. I think we're getting to that slightly quicker with AI because I think people were making big promises about AGI that have not delivered yet and probably won't deliver quite a few years yet, though apparently according to some, we were meant to have AGI today. And then again, people are starting to get their fingers burnt with LLMs and start to realize that the limitations of them what they're doing, you know, down to the fact that even the man on the street is starting to realize the hallucination limit of LLMs. And so you can't start to trust your entire business to something that is still not something you can properly trust, give you the right answer every time. So then we're seeing now well managed exercise caution. So people are investing in AI. There's no doubt about that. You know, you just have to look at the stock prices of Broadcom and video and others to realize that that's happening. And they are because people are investing heavily in infrastructure for training these models and that's what we, you know, selling the chips and the networking to do that. But in fact, in terms of using using those models and training those models and just right tuning those models and doing on-prem inference, I think we're now just starting to see those conversations happen broadly in the enterprise. And I've just, I commented yesterday in an internal call that it's literally the last six weeks every single one of my conversations with one of our big customers has involved AI. I'm prompted by us. They also, oh tell us about private AI. Tell us what's going because they're just starting to get to the point where we've done that bit of the hype curve. We've come down here and now we're like, okay, let's work on what we should be doing with it. So forming storming, norming. Yeah, and it is. And I think we've gone through that. We get you watch people go through that cycle. Some organizations are still going through that cycle of public cloud. They still think public cloud is cool and the thing to do. And you kind of like, can we just fast forward you 10 years to wherever one else is? And we're now realize you have to have a private cloud too. You're seeing other people accelerate through these hype cycles. So if anything, it's good because we're getting through these hype cycles faster. So latency, there's some industries where the cloud's never really an option. So things like telco's banks, manufacturers, you know, what's the clearest example you've seen where it just didn't work, where it was a massive failure, even though everyone committed to it? I don't think we've seen like a massive failure with everyone having committed to it and come back. I think there's always been industries which were never going to be able to go all the way, you know, both feet into the public cloud from the start. Usually around compliance, it tends to be about my data's got to be somewhere. I've got certain rules I need to meet. So you know, we see healthcare, for example, really sort of lagging in that space. Public sector, on the other hand, still think it's cool to go public cloud first. You can see the UK government announcing their new public cloud deals and we'll probably be on a public cloud. Whereas the larger enterprises, you know, the really big ones have sort of been through that cycle and now going through the, we've got to have private whatever. I don't think there's any, you know, in my mind, major examples of, you know, a whole industry that made a mistake. But what we do see more generically across the whole market is just people that have rushed in, poured loads of money into a project and then you go and see them. And the classic one was, I used to go and see someone and go, "Oh, okay. How's it going?" And they go, "Well, we've got this project where we've got to move, you know, 60% of our work clothes into public cloud by the end of next year." And you sort of go, why and they go I don't really know. I think some senior guy went and played golf with someone from a public cloud and we signed a deal and then he goes and he goes well here we are because I'm talking you know sometimes the sort you know minus one minus two from that layer the people had to do it and then he goes how's it going they go not very well you know it's it's not as simple to get workloads into the public cloud as people think it's not drag and drop it's you you have to refactor them in most cases you have to rewrite them and and whether they're IAS PAS or SASE makes even worse it you've got to go through quite a big change process to get in there and the reluctance to change I mean what what is it is it skills-based is it not invented here is it you know is it stock market 60% like you just referenced and we need to get to 60% because that's a number somebody stated somewhere yeah I think I always always amazes me the change thing right because I work an IT because I love the excitement of all the new tech you know technology guy that's my title right so I love the technology so for me the new gadgets the new features the new bits that we're doing that's so exciting to me and then I go and find people that work an IT that absolutely a bore change and you're like you are in completely the wrong industry why why are you even here if you don't like change you know and I understand there's a balance between risk and you know and advancing you got to be careful on that and there's regulations in some industries like finance and so on they've got they've got me to certain things but it's it's really quite similar across all industries when you go to that in terms of that balance of risk and it's so funny I go and sit with so many banks that say how we're unique we're special we're a bank you know no you're not you're like every other bank and to be fair you're not much different from some of the other industries we talked to and then you go and talk to defense and they go we're so special you'd never understand I bet you are words yeah I have conversations with every NATO member it's they're they all have very similar common challenges but I think that you know they they just hold themselves back with skills as you as you touched on it's a very very true one you know people just don't have the skill set in that technology but people hang on to what they know I mean you know I worked in the 90s for consultancies and it became very apparent to me that consultancies then were just going from one place the next copy and pasting what they'd done before I'll go into this because we've we've done a no-value network deployment before and so the answer to your problem is we're going to deploy no-value network oh brilliant fantastic I'll come and do that for you so you see a lot of that way people go I'm only going to go with what I know and you see that we're developers with a say oh I I must use Amazon's you know so and so service because that's the only one I know and you go well actually there's three different open source versions of that you can have one on-premise a version you can get in a seer there's one in Google you can talk to all the same API you can do the same thing but they just stuck hanging on to the one thing like they're cuddling the safe technology and you see that over do you think that's what it's got across do you think that'll get worse than as a technology is it expanding at an exponential rate I think this is where we have to come back to people have to start thinking much more about architecture we we come from a world of stacks right it used to be I'll go and buy the IBM stack and then I'll be fine you know I'll go and buy everything to pop no one's fight for buying IBM correct I'll buy the IBM stack or I'll go buy the Oracle stack or the Sun stack or whatever and that that'll do it for me and I think if anything what cloud has done for us is it started to stratify architectures and it started to separate things out into standards and APIs so it started to say okay well instead of writing to a stack what I'm going to do is I'm going to build things in Kubernetes I'm going to use Yamal files for my configuration we're going to do everything in containers that gives you a portability and so there's certain technology sort of strata that people need to look at you know look at these layers that you go into and so okay here's the level we operate out and so on and work up and that I think gives you that flexibility to use that new technology because they come in a daring to standards yeah there's thousands of different types of networks which is you can buy but they all use IP, BGP, OCEM, all that kind of stuff all those how you operate them still the same and so on and so on and so forth as you got the stack I think this is and AI will help us with that dramatically especially in the coding side in 2023 Broadcom bought VMware for 61 billion since then there's been a lot of things streamlined including one sort of flagship product VCF Cloud Foundation since then there's been a lot of mixed reaction how do we make sense of all the different reactions we're getting from people saying just completely different things yeah okay look there's I mean when the Broadcom acquisition was announced there was a broad range of reactions shall we say in the marketplace with people I think let's just talk about it from from you know the the only sector I've got which is you know my perspective being part of you know pre acquisition working through the acquisition and then working with the accomplishment of markets authority and others through that process and then post acquisition what we did let's look at VMware and then we'll sort of flip out and it come on to what you're talking about because it kind of it's an easy way to do it is VMware itself was an absolute hotbed of innovation right you know I was one of the leaders of the Office of the CTO you know there was the CTO and then myself and two others used to lead it and we literally were you know patents popping out every other day and cool new tech and cool new projects and we've got a load of stuff we could be incredibly proud of I think that one of the issues we kind of had at VMware was we almost never we we did a lot of blue sky stuff we very rarely said no to anything it was just oh let's try that let's try that and so we had quite a few products that you know we launched a bit kind of worked wrap them in did other stuff and so but the time we required we had I think it's depending on how you categorize it the way our R&D was organized was probably about somewhere between 9 and 13 business units building different bits of technology okay um sitting across the top of that was the office of the CTO we didn't have absolute authority we couldn't tell them what to do we were kind of more like a religious leader in sort of you know guiding them as to where they should be doing and doing it was a very heavy influence yeah very heavy influence yeah in broad terms of sect we don't have every technology company has some of that you trust me they have the there's there's technology cults in any business um so if you think about what we were trying to do let's we wind to sort of 2011 which is when I joined um VMware um I was brought on as part of a team which included the the future CEO Raggy Raggerum and you know the then CTO Steve Howard and a few others and we ended up I remember we we have an internal conference called radio which stood for the R&D Innovation Offsite which was pretty cool it was like the thing to go to is proper geek science fair it was awesome and then we I remember being in a bar in Monterey after that a bunch of us talking about this concept back then of software defined things and we said oh what about software defined data center and so you know you had software defined networking has been talked about software defined storage and so on and we said well the whole data center should be software defined because that's what essentially the cloud companies were doing what you know Amazon was doing verna was doing that and Eric and team were doing it at Google and what they were doing with Borg etc it's like okay fine so we decided what we should do is essentially take what we're doing with virtualization and build an operating system for the data center so the vision was this and this will become important later on is that you could basically rack up in your data center any tech you wanted right so I don't care as long as there's x86 processors who's got some storage attached to it in a network port we'll whack it in there and it's the building block of the data center we on the top we'll pour this layer of software defined data center this operating system for the data center which will then magically include that into the pool of resources and then you come to top and you just you know get access to things and every month you can go and buy some more hardware doesn't matter who you buy it from don't care because we just plug it in and just get it to the pool or whatever and so that's all the vision that's what we built so over the years we then had to say okay software defined networking we had kind of our own version of that but we acquired nice zero which became NSX which is you know literally you know open v switch it's the the de facto standard for virtualized networking now um christ of caram analyst and team in fact he I was really funny I bumped into him when I was at vmware but he was actually my lab supervisor when I was doing my degree in imperial college so I'm coming back I was like I'm sure I recognize you he invented v san amazing guy and so we built v san and then you know off the back of that we then invented a whole category of hypergivertian infrastructure that started as a project called project Marvin vmware which became evo rare and other things as well so we created this whole world around this this software defined data center except we were still shipping it as bits to customers we go okay well there's the storage bit and there's the networking bit there's the compute bit and here's a 400 page you know vmware very vvd we've kind of like how you should assemble it good luck that's nice brilliant there's some ps companies might be able to help you there and we could probably you know that's kind of what we're doing and we then built some integration on top of it which was vcf vmware cloud foundation so we had vcf3 then vcf4 and vcf5 which is where we got to just before the acquisition now the problem was that those 9 10 13 business units were all heading in slightly different directions they were all competing their own marketplace the storage be you was off competing in the storage world and the networking be you was off competing against Cisco etc another networking thing and the hypervisor be you was off you know making sure hypervisor was still number one etc and making sure we inventing cool things to go in there and so they'd all do their bits and they release them all at different times and if you ever anyone out there installed vcf3 4 or 5 you know the version numbers never matched up and the support matrix was always like some kind of you know impossible puzzle I'm sure it's going to pop up as the 1% question at one point and so you had this sort of and also it lagged so when you've you were running vcf and you came to me said oh there's a new version of vSphere out the new version of the hypervisor when can we install that so we've got to wait for the vcf team to catch up and love a plan and that frankly was the problem right so going through the acquisition Broadcom Hocktan the CEO Broadtongue who is an amazing man I've learned an awful lot from him came through and through the process acquisition we came to the conclusion that actually the product that VMware should be selling was vcf and what we were doing was we had you know thousands only really 9000 skews hundreds of bundles you could buy and it was all very very complex et cetera. Well, let's just simplify this You know if you got an Amazon you buy Public cloud right surely should be able to come to Broadcom and just go I want to buy private cloud. What's that? Well, that's this skew here And it's got everything you need to operate a private cloud near and it's this much and that's essentially what we did So we reorganized and this is the interesting thing we reorganized the whole of R&D Fliped it on its head. So now VCF leads What we do as an architecture all the components of that hypervisor network storage Orchestration automation et cetera. They all follow the roadmap of what's going on in VCF et cetera when it's released With VCF 9 came out all the version numbers matched that was more than just us changing numbers It was actually putting putting dev road maps in lockstep and then having a predictable future Edmaps because before if you come to me and said when are we going to release the next version of VCF? I go Not really sure I kind of have to do that and you know fill it in the waters Whereas now it's like I know it's nine months so we're released nine one you know nine months after nine and then you know about nine months after that and so we got a fairly regular cadence so From a product perspective the change has been absolutely Fabulous is all I can say I'm happy and I'm excited. I'm frankly still here Because I honestly think I'm getting to deliver and I really am with VCF 9 what we first tried to do back in that bar in 2011 That's the important part of the story that we had this vision and the operating system of the data center And so now I can honestly turn up to and go I really can give you a private cloud that you can put on prem that's better than the public cloud And that's VCF 9 and that's everything that we do in it now Granted that caused some upsets it changed how we interacted with partners it changed how we interacted with Service providers it changed what we did but interestingly Most people who go through that who've been through that journey come out of it and say God this needed to be done, you know a lot of the things we're doing was most of us again Wish we'd done this five years ago and we did a lot of it in a very short time So you know a lot of band there wasn't much breathing room was that it was just this is what we do in let's go with it Wasn't it well you imagine what it's like on the inside? In the outside it was like wow but from the inside it was it but it is you know I learned there's one word I would sum up from if you wanted to take what what's broadcon brought to all of this and what's the one thing You have its focus it's unbelievable focus focus on delivering the code on time, you know focus on delivering the operating You're disobrances based and it's just Such a focus that VM when ever had I think really and the people that dealt with the lovely company and we did all amazing Fantastic technology and all sorts of different ways but it was no no to clear focus and I think that's what's resonated with customers So when we've gone and spoken to customers it's very clear, you know hot tan stands on stage and says private cloud is it to your intro Right, you know and we all I now get to confidently going to customers and go private clouds it and they go Okay, and I got I've been saying that for the last 15 years, but now finally you might believe me and other technical people appreciates in the I guess the way that the stack all hangs up on the on tier numbers reference numbers and you know version numbers I think that the first number thing is is is interesting I think the interesting the the most important thing actually about what we've done in VCF 9 is around the operating model so We've tried it. It's much easier to deploy and much easier to operate. That's what we focused on Okay, so if you look at the release of VCF 9 look at the components There's a few cool things we've done you know in VSPIR we've got your memory tearing and we've got you know a few other Whistles and bells we put in there and each of the components has got some cool stuff that we've done in it There's no doubt we've got some new things in you know ESX 9 etc But all the real changes in that operations and automation layer that sits on top of it so yeah and Explore in London recently. I went and looked at the hands-on labs and after it I spoke to a few people and the sort of eyes were wide open like this is really cool This is really simple now it actually works. I've got the fleet level management stuff I was looking for I've got all the tools at my fingertips to be able to properly run a private cloud Rather than me feeling like I'm building it myself which is what we were doing before it feels like that Positivity's been lost then because obviously it's been quite a tough period of time Obviously the channel changes etc. Yeah, you know we've not heard massive positive news like you just explain Then is that something that we can work on? I think we I need to be telling more people the positive news don't I? Right for any kind of broadcast that we could do it on I'd say the um We know that the fortune 500 companies are going absolutely crazy for VCF right there. They are eating it up right so Three year five year ten year however many years deals they want seven and ten year deals regularly. Yes What why do you think that is why are they absolutely loving it and then you have these critics, you know who think it's not not good for them or it's not a good fit or Why do you think that is so The Large enterprises the fortune 500 etc the over 87% I think it's more than 90% Yeah, it's a big number. Yeah, you know that of our customers that have signed up all the strategic customers that I deal with It's all the same story Right the story was they've been staring at the public cloud for a while because that's what they were told to do because the public clouds They call shiny things they've been staring at the shiny like this rages all public cloud all And that's where the money's gone all the investment all the training all the skills etc And now they're realizing that they do need a private cloud that the public cloud isn't going to solve with their problems Whether it's you know cost, you know compliance or complexity those reasons why they need to do stuff on prem So then they're turning around and looking back at what they've got on prem and they're going That's not very good. So we'll jump in looking at a phrase you've literally let it sort of you know Wither on the vine or whatever you want whatever metaphor you want to come up with But also the interesting thing is it's also grown Because the on prem need didn't die away, but because you didn't manage it properly And I mean I go to so many customers large large enterprises and they've got you know three versions of the private cloud And this private cloud one two three and they've obviously got their company name Something X PC whatever it is with the name And realistically when you when you go and talk to them they suddenly realize actually we need to do this properly The challenges that on prem Typically they've never really followed the same rules that they did when they went to public cloud when you get to public cloud You're told that's it And you go actually I don't quite like that bit kind of that different note that's it That's what you do these you talk to these APIs it costs this much you use it Whereas on prem or the complexity comes because someone's gone Yeah, I really like this storage vendor And so all our storage is going to be by this storage vendor but that's not appropriate for all your work clothes Yeah, I don't care because that's what we're doing. I know that it's my favorite thing And we're going with this networking company because this network can come need blubber all that stuff and you end up with So what you end up with on prem is what people misguidedly I like to call best of breed So you don't talk to these large enterprises what they've got is they've got best of breed private clouds which don't work And they don't work because it's like trying to build a best of breed car If you want to go and build a best of breed car you go, yeah, I really really like I'm going to operate a fleet of taxis So obviously what's the thing you buy if you buy fleets taxis Mercedes right because everyone drives around Mercedes for fleets of taxis But they just happens to be and you're going your taxi firm that's got a 57 year history of working with BMW gearboxes So what's he's decided is actually we're going to buy some Mercedes taxis but we're going to run BMW gearboxes in them because that's what he knows and we've got really good expertise in middle of your gearboxes And there's someone else up there has gone actually I've got a really nice affinity for a particular type of engine from Ferrari from 1968 Which we really like and we run all our major systems of this for our engine So you end up with these taxis that quite frankly Oh, yeah, they've got really cool bits in them but put it together. It's disgusting And then you turn back up to Mercedes and go can you fix this and they look and go what the hell have you done I can't fix this in your sense And that's basically what you see on on prem in most people whereas now it's it's absolutely beautiful because I'm now getting CIOs that I sit with And they're CTOs and whatever you want to call the levels underneath that and it varies massively saying to me I'm dictating that we're going to take VCF from you And we're going to run it as you say your architecture your components And if you catch any of my team ripping it apart and trying to swap the gearbox and swap the engine and swap whatever Tell me because that we're not going to get in that problem again So here's that's the real difference what they the big changes they're going okay fine I know I need to have a private cloud on prem But I need to run it the same way as a public cloud if it's going to be the same as public cloud It's going to be cheaper than a public cloud and better than a public cloud I need to listen to the people that are telling I mean look You don't buy servers from Dell and then rip bits out of them and put other bits in them do you buy and manage you know That's what you buy the H whatever it is so just to go a layer up private cloud buy that software from someone and just run it And I'd say that's the that's the real difference between the old VMware as well because as you said used to be sort of like a Parts company and now you're actually selling the car. Yeah completely. Yeah, you know So it is crazy to think I wanted to because one of the metaphors were used back in the day when we looked at the software defined data center We looked at those public clouds and we said okay, you know look at Google for example Google was the equivalent of a formula wanting you know when it came to running a public cloud They had the most expensive best engineers in the world They were running one cloud that did one particular thing which was you know running you know their platform and Well, there's kind of things and if you turn around and you were co-op banking Manchester and you wanted to do it You didn't have the f1 engineers you didn't have all that kind of stuff so you couldn't do that you couldn't do that on prem So what I said was we need to almost be like the Toyota back then of of Of Carmen if actually we can come to us and you can come and buy You know because not everyone wants an f1 car But what they want is that car that's going to be do what they want is easy supportable repeatable scale model all that kind of stuff and That's so yeah, it comes up to the same thing. I Guess that I'd take was swiftly into some customers think that VCF restricts their choice and a lot of customers like to pick you know what what they want in their stack What do you what do you say to that? so The leaders in our industry from a IT leadership perspective have two words they want to talk about standardization and simplification. Every CIO has a mantra around standardization simplification. So a lot of their challenges come from that over complication before and almost giving people too many choice. So depending on how you look at it, yeah, okay, you could say, "Oh, we're selecting NSX, but you're selecting NSX networking as opposed to what?" And other hardware provider that's going to tell you hardware-based networking. Are you selecting V-SAN, which allows you to attach any kind of storage you like to it, basically, to any type of server? No, actually, what I want to do is I want to go to this other storage company that is very prescriptive about how they put their storage in. So actually, when you think about it, what we're trying to do is give people more choice. So in that VCF layer, you get a choice. You can run on any, you know, the storage is V-SAN is awesome. You can attach loads of different types of storage to it. You can buy different, different, different people at works. NSX works on any white box flat layer two network you want. Just go and buy yourself some flat layer two network. We do it internally. It will come, it's all white box stuff running sonic. It's phenomenal. We're not tied to any particular networking operator. We have more choice, right? Storage, we have more choice. Processors, well, fantastic, right? Any X86, but now, of course, we support GPUs as well. So we support multiple different types of GPUs. And going forward, you're going to see XPUs, TPUs, DPUs, and others. And that's on our road map. So you'll have even more choice of what runs in this heterogeneous data center. So actually, what you're getting is you're getting a layer of standardization and simplification that supports all these lovely protocols northbound like Kubernetes, et cetera. But southbound allows you an amazing amount of choice. So you're not tied to any particular hardware vendor, any particular technology. You get that real flexibility and running in hundreds of public clouds. So it's more choice. It's just different people have a different lens. So we touched on data sovereignty earlier on. So who controls the data and under which laws? So we've obviously seen Police Scotland, ask Microsoft, and they couldn't answer the question where the data was being processed. We've seen the French Senate take Microsoft to court because they couldn't guarantee that French data held in Paris, stayed in Paris so I didn't end up in the US. So how's VMware going to deliver? What's the alternative? Yeah. Sovereignty is about caring about where your data is and who has access to your data. There's a couple of things, right? So where is my data and who has access to it? Is it only me or could other people do it? The big fly in the ointment with all of these things, the examples you just came up with is the US Cloud Act, which means that the operator of a cloud, if there a US entity, doesn't matter where that data is in the world, if there a US entity, then the US government can ask and get access to data anywhere in the world, which is why these hyperskirts do that. So if you're taking a cloud that's operated as the service provider by a US entity, you're never going to solve that problem. So you'll see a lot of fantastic marketing by cloud providers where they start to talk about trusted clouds and other things or whatever. But the problem is if it's still a cloud, it doesn't matter if it's operating out of a data center based in the UK, as the police Scotland found another, the potential is there because they're the operator of a US company that they could come and the US government could come in and do that. Not saying they would, but they could. They could, right? So that's a source of major concern, both from private sector, some people are concerned about that. We see it massive concerning public sector, critical infrastructure, all those kind of things people are concerned about what they do with that. So what's the answer? Well, the answer is you don't use a service that's operated by an American entity or a Chinese entity or any other foreign entity that you don't want to trust. So what you do is you operate it yourself or you get a UK based entity in the context what we're talking about here to do that. So you get a UK cloud provider to stand up a UK cloud in UK data centers using UK staff with a UK legal entity paying UK tax as well, which is quite nice, right? And then we will provide the software for them to operate that public cloud. So that's our answer is we say, okay, fine, well, we will give you the software to operate the cloud. We have nothing to do with it. We're just a provider of the software. Much like you're going to go and buy the x86 chips from some other foreign entity or whatever. But once it's up and running, it's your control. The benefit of that though is also a lot of people are finding is because all these public clouds are running VCF and it's exactly and directly compatible with the VCF they're running on prem in their private cloud, they can freely smooth stuff to and from. So we're seeing a lot of people realize that they can run a very powerful sovereign enterprise, both on prem and still in public cloud providers, but the right public cloud provider doing the right service. And is the AI playing in this because you read about anything you put in chat GPT that ownership goes off to wherever? Yeah, there's some fantastic horror stories aren't there always will be? Yeah, you know, people getting source code popping up that other people are asking us, yeah, so of course, and that's the other, you know, that's the leading concern that I'm getting when I'm talking to IT leaders and business leaders is that they can get into AI, but their data stays their data. And so that's where we came up with the concept of private AI. And we talked quite openly and often about private AI, but the concept of that is that really we support AI and we have an offering that's now part of VCF, which we announced in export, it comes free, you know, included in VCF 9 is the private AI, a pay which sits in there. And what that does is that enables you to support AI workloads on VCF, whether it's on prem or in that sovereign cloud. So now you can go actually what I'd really do is what's really powerful for me is I'd love to have an LLM that's pointed out, every contract I've ever written, when my teams are writing new contracts, they can use that as an assistant, but I want to know that all that data stays within my direct control, you can do that with private AI from VMWR.com. But it goes beyond that because people are now saying, okay, well, actually there's permissioning levels in that. So we can do so AI has a concept of RAC, which is a retrieval augmented, which is essentially a database that sits alongside, right? Now, we can do a permissioning level thing within that, so you can have multiple different levels of permissions within those RAC databases. So if you're a government entity, for example, you've got, you know, restricted, confidential secret, top secret, we can have four different databases, all put it into the same LLM, and depending on when you come to it, if you're only allowed to access up to level two, you'll only get the LLM answering from the data and level two, but if you're access someone from level four, we'll get answer from all of them. So we've got again, that granularity permission in there too. So it's, and the other thing is cost. It's not just about people being concerned about the AI, you know, having access to data that don't have access in the public cloud, but it's cost. We're running AI typically on-prem, but a third of the cost in an inference basis, then it costs in a public cloud. So people are not doing it just because it's, you know, data protection is also cheaper in most cases. You've said before running a proper cloud. Yeah. And should mean thinking like a cloud provider. Yeah. How does VCF change this? So it really changes it in both the deployment and the organization models. So we have a, we have a, we take our customers to think of the PCM, which is about private cloud maturity, right? And that's really helping customers change their organization as well as change their technology. So, you know, I could go and drop VCF 9 into you tomorrow, but you'll, you'll never make the most of it if you don't change your organization to think like a cloud provider. So more importantly, we got rid of all our silos to build VCF 9 within our organization. If you think about all those different views, and we've created one, you know, thing focused on VCF, so we need our customers' organizations to do that. And it helps them do that because it's got all the bits and pieces in there to help people be a bit more organized around doing that in terms of the dashboards and the controls and the compliance and the drift commandagement and all the fleet management stuff that we got in there enables you to do incredible things at scale that you've never been able to do before. And I think that's the real thing. So let's talk about scale because that's really exciting to me. So Broadcom, when they're acquired VMware, VMware had 45,000 physical hosts for 40,000 employees, which is an interesting one for a virtualization company, right? You'd think that. And over the process of the consolidation of the Broadcom existing infrastructure and the VMware infrastructure, we now have an environment that supports 27 unique individual businesses because that's how Perlcom operates, some of which literally are like Pepsi and Coke in terms of the chip business they can't talk to each other, etc. So they're entirely separate. Some of them are semantic, testing malware. So oh my god, you want that stuff properly separated. So we run one central cloud platform for 27 business units, okay, for a $1.6 trillion company. So there's seven, three eight largest companies coming in the world. Only going up. Well, and only going up. Yeah. On 15,000 physical hosts, we went from 42 to seven data centers in 14 months. And we went from two and a half thousand people managing our private cloud to 68. So how does VCF9 help? That's how it helps. We can get you down from two and a half thousand staff to 68 staff running a cloud that 15,000 physical hosts actually supporting more than before across 27 discrete business units that run globally in a $1.7 trillion business. So when I walk into a bank that's 55,000 people know I T department, I'm sort of a bit RCO can't believe it. It's a bit of a gas. It's like there's probably 10 times too many and even then 10 times too many. So is this your total cost of ownership example then? Yeah, it's one of them. It's one of them. It's one of them. Our internal Brawkommon Brawkomm and people like, oh yeah, but no, our CIO, because the way the Hocktown Munstink had to pay for the software. So you didn't get the software for free because you don't get it for free. So you had to pay for the software. But more importantly, our CIO directly reports to Hocktown and has the direct support of him. But if you want to talk about standardization simplification, oh wow, the standardization simplification in Brawkommon, you don't get to say, I want that. It's like, no, what you're given is this. That's what you're going to work out and operate like that. So it's very lean, very very slick on operations model in IT and also multi-skilled and that's the other important things you talk about how do people organizations change? Go and talk to Alan Davis in our CEO. He'll tell you there's no one who's just a networking person They're networking and storage their compute and networking. They're all multi-skilled because it's I don't need single-skilled people because you've got to think multi-skilled in this world So if you are think you're moving on in your career and you're just thinking about I'm the networking guy You're not going to go very far in the future world of building integrated private clouds So I was at the channel for a couple of years certainly when a VMware was purchased by Brocom The misconception from afar was that it was there was a brutal Approach to the channel prices went off people got terminated The price increases versus a total cost of ownership example seem to be a misnomer How how can we address that how can you guys address that for us because obviously when you know we're talking to our customers You know everyone is complaining about the cost increases, but actually it sounds like it's far more of an offering to for them to take up Yeah, so the problem we are the problem you've got generally have is we're going to talk to a customer and the customer is Mainly seeing us as someone who provides their hypervisor We're not doing networking for them. We might be doing a bit of storage for them So when you start going to say okay, well here's VCF. They're almost like well That's way more expensive than I'm paying. It's like yeah, but you're getting so much more now of course That's a bit of a weak argument right You know because you're getting so much more but actually if you do want to make the most of it You will find that you will lower the total loss to ownership is but it's going through that journey with people I we've just been through it with a customer Very very very large business and they Sort of came to us and said here's our architecture and we walked in and we halved the total cost of running that By deploying VCF and they said here's our infrastructure to run VCF on that would cost this much and you're going Well, no, it wouldn't wherever you got that number from will be spoke to some other people in other companies We didn't talk to the people selling it to you so come and have the conversation with us and I think that's a lot of Here's saying rumours going around and we'll go often sit across the table from the The top people in a very large bank going oh renewal. We know it's going to cost this much You're like where do you make that number up from? Well, we kind of guessed right? Yeah, I was going to say if you for instance if you had a CIO come up to you and say you know where we're moving away from VMware because your price is extortionate What would you turn around and say to them? I would be a conversation about what they're trying to achieve with their IT Literally where are you going and where do you because like I say the They probably got what we do framed in the wrong context they probably go well I'm looking at different hypervisors and think well if you think all we are as a hypervisor then you've got us wrong Right, you know, I'm going to switch away go and use you know some other random thing Well then fine you're probably not the right person for VCF in that in that context anyway if you're just looking in that way But if you're saying actually what I want to do is I want to build a proper private cloud A real you know a real private cloud that operates better and cheaper than a public cloud than then we'll have a conversation and We we went into a to finish the example I went to this customer and they sort of said you know We've priced up what we think was it will cost us to renew and you went well Yeah, but if we actually went look to your environment I can half the number of service you've got based on your current architecture And if we half the number of service you've got over the next three years and look at then how you won't need to buy more service and everything else I've had that and look at the architecture and oh look let's change how your networking is going to be done because you don't know the physical networking I know all those low physical low balances you've got or you can do that with avian software and blah blah blah blah We ended up with an architecture that ended up including the cost of VCF 9 was less than they were paying today just for us You know, so it's like your total cost is is going to be much better but the problem is people are still doing their best to breathe thinking Right, so it's like okay well, I've come to you and you've just told me that the price of my engine is going to double And that's going to make my car really expensive like well if you thought about maybe trying a whole different car That's better modern faster integrated does everything for you know, I suppose that's really the way to think about it Is the example you shared with us is that I want to create my architecture my infrastructure to be a cost-efficient solution or is that to be at the leading edge of whatever they do in their business unit It is there a difference between I want to be driving change whether it be online selling you know like any website you portal or is it just you know we know what we're good at we're sticking to our knitting and we're just going to make it the most efficient Back office system we can have what what where do you play in that are you at the leading edge or just the real cost management cost Contelment and say what do you think I'm going to answer or say all of them? Yeah, of course. I'm going to say both right so if you go into the bank for example They have two major projects in any bank run the bank change the bank right? It's all you do run the bank change about run the bank change the bank for years They've been trying to bring them together the difficult thing for us is we can do both We're going to go in and help them lower the cost they run the bank whilst changing the bank But the problem is that it's two different people to talk to you know It's kind of a challenge right and so that's where their organizations hold them back But yes, we are obviously Helping them down so Finops is a huge conversation right for large enterprises Finops is a huge conversation We've got a whole bunch of stuff in VCF 9 that helps with Finops in costing prediction all that kind of stuff which makes a big difference So yes, we're doing that but at the same time we're deploying and this is the other thing I went to someone else after recently said you know He saw everything that we did on prem as legacy and saw everything in the public cloud as being future and modern It was like well no I can do everything you want to do in the public cloud that you think is future and modern in terms of application architecture services You know all the way you'd want to build stuff But I can do it on prem and I can do it probably a bit faster and a bit cheaper for you But they hadn't thought like that. So you've got to do both yeah because the answer you have to take people down that journey Both and that's difficult because you're trying to get people to change a lot of things at the same time And people aren't good at changing a lot of things people aren't good at a lot of variables You know they just can't change a lot of things at the same time So us going and saying actually we've got you know you're going to change how you operate you're going to change what you operate You're going to change the architecture everything else but We're getting really good at going through that process with the customers Certainly the strategic and the you know the top-end area with our consulting and then with our partners The consulting partners that we've got that are going at and doing that I think that's really where they're going to have to help is it's not just about tech it's about operations models And it's so important that people understand that that it is about first thing I said people in process You've got to get the people in process right because that is what's going to hold you back So the reason why it's going to be more expensive is probably not ask it's your people in process and the way that you're going to do it is going to make it more expensive Your architecture is probably going to make it more expensive I remember customer recently and talked to them I said them I found a new way to lower the number of cores you're using and they looked at me and what are you doing? I said well I've just found a way to literally reduce your core usage by a third And they were sort of literally and well why would you do that because that means we'd have to pay you less I said yeah, but I'm not here for this week. I'm here for the next 10 years. Yeah We're here for a long-term partnership with you So I want you to see you're getting value out of that. I don't want to come in and say oh yeah We've we've opted to speak thing up and sell you load a harbor and walk away as some other organizations would So that's went and that those conversations when you start talking about TCO you start talking about long-term value you start talking about Evolving their people in process being competitive new technologies having a choice in the platform going forwards of what runs on it and what they can run on it That's when someone goes I'm in with you for seven years. I'm in with you for ten years What's the deal? How do we do and those are the conversations we're having it's really it's exciting because we never really had those conversations before at the Endware because you say we were always a bit kind of a parts seller, you know You're always going in talking to bits of the business now you know if you start talking You know as with as broad-come has been we're having more mature conversations and who would you credit with that focus from like the multi-layered multi-offer into The single point of VCF who's the you know who's that vision is that hot time? Yeah, I tell you it was hot time. I mean honestly it was hot. Can you saw that the He saw what the market opportunity was he saw what people were thinking in the customer? I mean, I mean I'm not kidding Hot-tann can pick up the phone to pretty much most CEOs and CEOs in the world and Have a chat with them. It's quite amazing to be honest and that's what he does a lot is you spend his time with those people He spends more time with the I've never seen a CEO spend more time with customers ever Than hot-tann certainly not in our business. I mean literally he's all the time he's out seeing a largest customers Whether they're coming to see him in Palo Alto who's out on tour seeing them and so when you talk to him His finger is absolutely on the button as to what our customers want what they need what's causing them problems You know all those kind of things and trust me are Our product teams feel it they feel it they're talking about the word focus. It comes direct from mr. Hot-tann himself So yeah, that vision that simplicity that it's it's and it's been Frankly, it's been you know It's been we've done it quickly don't get me wrong and yeah There's been bumps along the road as I think you said but the focus has just been you know Relentless, but it's been refreshing. I found it. It's very very good. Yeah, honestly And who would you give credit for the pricing as I said when I was out the channel The miss no more is it is all become far more expensive with doubled in price yeah Obviously, you you know you market cap and the perks rise going from 60 plus billion to a wean to mid 70s 80 billion Now is it something like that maybe even higher what for the other we don't call the VMware bit out But I know that a stock price is more than doubled since your position So who would you credit with that sort of pricing strategy then because there's an accountant somewhere Who come on Joe take the credit this I'm the chief technology officer. I'm not the chief pricing or modeling or licensing officer But as a massive play to to do what the pricing appeared to do from outside Mm-hmm. I'll use the phrase double it and you know the businesses continue to grow exponentially In a fairly flat market. Well, I it's it's and not been on AI as well by sorry to cut crushing No, no, no, no, I don't you know I'd all credit. I think it's been it. It's been a very well-executed journey No one name you're going to give credit to No, I can't remember. Hawk Town again. I swear really Or Rose leave that yeah If I think we should quickly touch on automation and maybe a little bit of AI wrapped around VCF. I was at Explore, obviously we met Explore in London and they were going on about automation and how this is going to change for the IT teams out there. Yeah, I think so we've had automation in the cloud platform for a very long time since the acquisition of dynamic ops many, many years ago, which ultimately became V Realize and then RIA. So it's been for various things. But essentially what you're looking at that automation and operations feature, which is VROPS and sort of come together. I gave a keynote in 2011 to VMug in the UK in the National Motorized, Motorcycle Museum in Birmingham and it was Automate or Die was kind of my mantra because it was again, those people out there that were doing their components and I said if you're specifying scousy cable lengths you're probably not going to be around much longer, right? You know, the trick is going to be an automation. The trick is going to be in scale. The focus that we put in the last 18 months or more to complete what we've been doing for a few years but accelerate it massively around those fleet management features is just huge and I think it's what everyone gets most impressed by and it's automation of a couple of things. It's automation for two audiences because you've got to automate for the IT admin but also automate for the customer, the developer. So the fact that we have this, you can see the Explor-Kino, there's a demo for the IT admin and there's a demo for the developer because we've got to talk to both these audiences and the IT admin audience is all about helping you manage risk and manage the security and set guard rails and make sure that you can still do things at speed and make sure that for the developer this is of a fast flexible environment that is secure. So when you look at what we're doing automation there it is those fleet level management things. The stuff we put in there like the live upgrade, the live upgrade features. No downtime. So a non-evacuation upgrade of a host, unbelievable. So cool. That was something I saw previewed as an idea at radio, the R&D conference, probably about seven or eight years ago. We'll try and do this and use the V-motion technology and other bits to do what we're doing there. So there's some really really cool stuff in there and the bits in terms of how we're doing the configuration drift management and all the other bits and technology that we've got in the automation layer. It means that someone's sitting as the console for an IT admin should be essentially not having to do as much as they did before in terms of trying to pull bits and pieces together to make this all work. And yes of course we're putting AI assistance in there as well to make it easier so you can ask natural language what's up with this and what's the best way to do this. We don't have remediation in there yet because we don't quite trust the AI to do that yet. But at least give us some hints and tips on what you should be doing and then go look and be able to do that and bring that in and then integrating and that's the other thing is the integration piece, so integrating V-defend, the Brimmer Live Recovery, the ransomware recovery which is a huge topic right now. All of those things being much much easier and integrated and managed from one console in essence is much easier. I'm never going to say single pane of glass because it's the horrible thing but bringing it together as much as possible people. And the lovely thing about it is when people see that console you've got a GUI, you've got a CLI now and you've got the API as well. So if you want to do it through the GUI you can click and point. If you want to do it through command line you can go straight to a terminal and start doing it through a command line. If you want to whack the ammo files at it you can do it through that as well. It doesn't really matter. And in fact the nice thing is as you're building something in the developer interface it's actually building those out for you as well so you can flick between the two. So I get very confused very quickly sometimes in my home lab playing about them and yes I do have one playing about between some of these things. But yeah it's phenomenal. So for me you know when I look at it the ease of use is just huge because if you're going to get 68 people to operate a private cloud for a huge company you've got to have that level of automation though you can't use this without it. And it doesn't really make much sense when you have CIOs complaining saying it's not going to work here it's not going to do that I want it to do I want my total cost of ownership to go down. You've got to think they must not understand the product completely. You know where are they getting this information from. I would say for IT leaders who are watching this now and they feel maybe left behind by the shift what would you say to them. Automate or die. Bring it back. We need to get down a shirt you can find them link below. Yeah I got a commission on that as well please thank you very much. No seriously that's it if you're going to do this stuff at scale you've got to automate. If you want the simplicity you want the standardisation you've got to automate. And I think the problem is the things that stop people automating and people see themselves being automated out of a job. I think that's part of the problem. You get to an organisation and say oh actually if I do all that automated then what do I do all day. The answer is you move up a level and you control more things and you get bigger right. You don't sit there and watch yourself get automated away. You move yourself up and become the person that does that. It's interesting because that's going to that's a continual thing in technology right. You know continual thing. You know yeah I was back in the 90s when I was running at Compact Proline 4500s and getting excited about raid controllers and things and what first and a firmware it was. That's fine but no one cares about that stuff anymore because it's just black boxy stuff we just whack it and it's fine. And a friend's tour of AI I had a friend's son who's just graduated in video editing. He's got degrees and three year degree in video editing and I said so what are you doing and he said this is about two years ago and he sort of came to me so I said what's your plan? He said well basically I've set up this website and what people do is they upload clips and they tell me how they want the metadata together and then I send it back to them and they say can you change this bit into that bit and I basically do an online free video editing service. I went using an AI tools to do that. No no no no no I said well okay well you need to start using AI tools to do that because otherwise what you're doing is going to be replaced by AI because that's what you're describing is exactly what AI is saying it's going to do is you should be able to upload a load of videos to an AI engine and then mash it together for you and then you go no move this bit change up whatever so I've just replaced you what you need to be is you need to be some person using the AI tools and it's the same for developers and people think oh we're going to get rid of all of the developers because we're going to use AI now actually that's not happened at all what's happening is you're now getting developers augmented AI and so on and so forth so actually all this technology augments people so when I say automate or die it very much is augment yourself with these automation tools or become part of the change be the visionary that says I'm going to be the one that automates private cloud I'm going to build a private cloud my company that's what I'm going to do not I'm going to fight it back because I've got this particular network vendor or this storage vendor written through me and I'm never going to change so we call this podcast state of systems for a reason and we always have a closing question so on a scale of one to ten how confident are you that private cloud is here to stay I don't think the scale reaches where we need to be does it well you know we just have spinal tap too coming out right so I'm going to have to go to 11 on I mean look I'm my career is better on it you know the stock market's betting on it too and I think you know you can just see for my customers private cloud is here to stay people are going to need it so it's definitely an 11 without question thank you so we did want to dig deep as well on the bro come just released the private cloud outlook 25 the maturity advantage please tell us more so we went out and did a survey of hundreds of our customers over 200 in Europe alone I think it's 200 in the UK actually there's much much more than that hundreds of our customers coming out to properly asked them you know the question about you know what are you doing about private cloud what are you doing public cloud are you you know why are you coming back on a prayer with private cloud why you're repatriating how many you're repatriating there's some fantastic stats in there about exactly what people are doing but the overwhelming trend is people have landed on the fact that and if you need that data support to go and you know make your own decisions or make your own justifications for why you need to have a private cloud basically people have decided it's right workload right place they're going to have to have public cloud they're going to have private cloud and think any private cloud let's do it right and we will put a link to the private cloud outlook in the show notes as well thank you so much Joe Bagley for joining us today it's been fun you [BLANK_AUDIO]