Hello and welcome back to Supercloud 2 where we examine the intersection of cloud and data in the 2020s. My name is Dave Vellante. Our Supercloud panel, our power panel is back. Maribel Lopez is the founder and principal analyst at Lopez Research. Sanjeev Mohan is former Gartner analyst and principal at Sanjeev Mohan. And Keith Townsend is the CTO advisor. Folks, welcome back and thanks for your participation today. Good to see you. >> Okay, great. >> Great to see you. >> Thanks. Let me start, Maribel, with you. Bob Muglia, we had a conversation as part of Supercloud the other day. And he said, "Dave, I like the work, you got to simplify this a little bit." So he said, quote, "A Supercloud is a platform." He said, "Think of it as a platform that provides programmatically consistent services hosted on heterogeneous cloud providers." And then Nelu Mihai said, "Well, wait a minute. This is just going to create more stove pipes. We need more standards in an architecture," which is kind of what Berkeley Sky Computing initiative is all about. So there's a sort of a debate going on. Is supercloud an architecture, a platform? Or maybe it's just another buzzword. Maribel, do you have a thought on this? >> Well, the easy answer would be to say it's just a buzzword. And then we could just kill the conversation and be done with it. But I think the term, it's more than that, right? The term actually isn't new. You can go back to at least 2016 and find references to supercloud in Cornell University or assist in other documents. So, having said this, I think we've been talking about Supercloud for a while, so I assume it's more than just a fancy buzzword. But I think it really speaks to that undeniable trend of moving towards an abstraction layer to deal with the chaos of what we consider managing multiple public and private clouds today, right? So one definition of the technology platform speaks to a set of services that allows companies to build and run that technology smoothly without worrying about the underlying infrastructure, which really gets back to something that Bob said. And some of the question is where that lives. And you could call that an abstraction layer. You could call it cross-cloud services, hybrid cloud management. So I see momentum there, like legitimate momentum with enterprise IT buyers that are trying to deal with the fact that they have multiple clouds now. So where I think we're moving is trying to define what are the specific attributes and frameworks of that that would make it so that it could be consistent across clouds. What is that layer? And maybe that's what the supercloud is. But one of the things I struggle with with supercloud is. What are we really trying to do here? Are we trying to create differentiated services in the supercloud layer? Is a supercloud just another variant of what AWS, GCP, or others do? You spoken to Walmart about its cloud native platform, and that's an example of somebody deciding to do it themselves because they need to deal with this today and not wait for some big standards thing to happen. So whatever it is, I do think it's something. I think we're trying to maybe create an architecture out of it would be a better way of saying it so that it does get to those set of principles, but it also needs to be edge aware. I think whenever we talk about supercloud, we're always talking about like the big centralized cloud. And I think we need to think about all the distributed clouds that we're looking at in edge as well. So that might be one of the ways that supercloud evolves. >> So thank you, Maribel. Keith, Brian Gracely, Gracely's law, things kind of repeat themselves. We've seen it all before. And so what Muglia brought to the forefront is this idea of a platform where the platform provider is really responsible for the architecture. Of course, the drawback is then you get a a bunch of stove pipes architectures. But practically speaking, that's kind of the way the industry has always evolved, right? >> So if we look at this from the practitioner's perspective and we talk about platforms, traditionally vendors have provided the platforms for us, whether it's distribution of lineage managed by or provided by Red Hat, Windows, servers, .NET, databases, Oracle. We think of those as platforms, things that are fundamental we can build on top. Supercloud isn't today that. It is a framework or idea, kind of a visionary goal to get to a point that we can have a platform or a framework. But what we're seeing repeated throughout the industry in customers, whether it's the Walmarts that's kind of supersized the idea of supercloud, or if it's regular end user organizations that are coming out with platform groups, groups who normalize cloud native infrastructure, AWS multi-cloud, VMware resources to look like one thing internally to their developers. We're seeing this trend that there's a desire for a platform that provides the capabilities of a supercloud. >> Thank you for that. Sanjeev, we often use Snowflake as a supercloud example, and now would presumably would be a platform with an architecture that's determined by the vendor. Maybe Databricks is pushing for a more open architecture, maybe more of that nirvana that we were talking about before to solve for supercloud. But regardless, the practitioner discussions show. At least currently, there's not a lot of cross-cloud data sharing. I think it could be a killer use case, egress charges or a barrier. But how do you see it? Will that change? Will we hide that underlying complexity and start sharing data across cloud? Is that something that you think Snowflake or others will be able to achieve? >> So I think we are already starting to see some of that happen. Snowflake is definitely one example that gets cited a lot. But even we don't talk about MongoDB in this like, but you could have a MongoDB cluster, for instance, with nodes sitting in different cloud providers. So there are companies that are starting to do it. The advantage that these companies have, let's take Snowflake as an example, it's a centralized proprietary platform. And they are building the capabilities that are needed for supercloud. So they're building things like you can push down your data transformations. They have the entire security and privacy suite. Data ops, they're adding those capabilities. And if I'm not mistaken, it'll be very soon, we will see them offer data observability. So it's all works great as long as you are in one platform. And if you want resilience, then Snowflake, Supercloud, great example. But if your primary goal is to choose the most cost-effective service irrespective of which cloud it sits in, then things start falling sideways. For example, I may be a very big Snowflake user. And I like Snowflake's resilience. I can move from one cloud to another cloud. Snowflake does it for me. But what if I want to train a very large model? Maybe Databricks is a better platform for that. So how do I do move my workload from one platform to another platform? That tooling does not exist. So we need server hybrid, cross-cloud, data ops platform. Walmart has done a great job, but they built it by themselves. Not every company is Walmart. Like Maribel and Keith said, we need standards, we need reference architectures, we need some sort of a cost control. I was just reading recently, Accenture has been public about their AWS bill. Every time they get the bill is tens of millions of lines, tens of millions 'cause there are over thousand teams using AWS. If we have not been able to corral a usage of a single cloud, now we're talking about supercloud, we've got multiple clouds, and hybrid, on-prem, and edge. So till we've got some cross-platform tooling in place, I think this will still take quite some time for it to take shape. >> It's interesting. Maribel, Walmart would tell you that their on-prem infrastructure is cheaper to run than the stuff in the cloud. but at the same time, they want the flexibility and the resiliency of their three-legged stool model. So the point as Sanjeev was making about hybrid. It's an interesting balance, isn't it, between getting your lowest cost and at the same time having best of breed and scale? >> It's basically what you're trying to optimize for, as you said, right? And by the way, to the earlier point, not everybody is at Walmart's scale, so it's not actually cheaper for everybody to have the purchasing power to make the cloud cheaper to have it on-prem. But I think what you see almost every company, large or small, moving towards is this concept of like, where do I find the agility? And is the agility in building the infrastructure for me? And typically, the thing that gives you outside advantage as an organization is not how you constructed your cloud computing infrastructure. It might be how you structured your data analytics as an example, which cloud is related to that. But how do you marry those two things? And getting back to sort of Sanjeev's point. We're in a real struggle now where one hand we want to have best of breed services and on the other hand we want it to be really easy to manage, secure, do data governance. And those two things are really at odds with each other right now. So if you want all the knobs and switches of a service like geospatial analytics and big query, you're going to have to use Google tools, right? Whereas if you want visibility across all the clouds for your application of state and understand the security and governance of that, you're kind of looking for something that's more cross-cloud tooling at that point. But whenever you talk to somebody about cross-cloud tooling, they look at you like that's not really possible. So it's a very interesting time in the market. Now, we're kind of layering this concept of supercloud on it. And some people think supercloud's about basically multi-cloud tooling, and some people think it's about a whole new architectural stack. So we're just not there yet. But it's not all about cost. I mean, cloud has not been about cost for a very, very long time. Cloud has been about how do you really make the most of your data. And this gets back to cross-cloud services like Snowflake. Why did they even exist? They existed because we had data everywhere, but we need to treat data as a unified object so that we can analyze it and get insight from it. And so that's where some of the benefit of these cross-cloud services are moving today. Still a long way to go, though, Dave. >> Keith, I reached out to my friends at ETR given the macro headwinds, And you're right, Maribel, cloud hasn't really been about just about cost savings. But I reached out to the ETR, guys, what's your data show in terms of how customers are dealing with the economic headwinds? And they said, by far, their number one strategy to cut cost is consolidating redundant vendors. And a distant second, but still notable was optimizing cloud costs. Maybe using reserve instances, or using more volume buying. Nowhere in there. And I asked them to, "Could you go look and see if you can find it?" Do we see repatriation? And you hear this a lot. You hear people whispering as analysts, "You better look into that repatriation trend." It's pretty big. You can't find it. But some of the Walmarts in the world, maybe even not repatriating, but they maybe have better cost structure on-prem. Keith, what are you seeing from the practitioners that you talk to in terms of how they're dealing with these headwinds? >> Yeah, I just got into a conversation about this just this morning with (indistinct) who is an analyst over at GigaHome. He's reading the same headlines. Repatriation is happening at large scale. I think this is kind of, we have these quiet terms now. We have quiet quitting, we have quiet hiring. I think we have quiet repatriation. Most people haven't done away with their data centers. They're still there. Whether they're completely on-premises data centers, and they own assets, or they're partnerships with QTX, Equinix, et cetera, they have these private cloud resources. What I'm seeing practically is a rebalancing of workloads. Do I really need to pay AWS for this instance of SAP that's on 24 hours a day versus just having it on-prem, moving it back to my data center? I've talked to quite a few customers who were early on to moving their static SAP workloads onto the public cloud, and they simply moved them back. Surprising, I was at VMware Explore. And we can talk about this a little bit later on. But our customers, net new, not a lot that were born in the cloud. And they get to this point where their workloads are static. And they look at something like a Kubernetes, or a OpenShift, or VMware Tanzu. And they ask the question, "Do I need the scalability of cloud?" I might consider being a net new VMware customer to deliver this base capability. So are we seeing repatriation as the number one reason? No, I think internal IT operations are just naturally come to this realization. Hey, I have these resources on premises. The private cloud technologies have moved far along enough that I can just simply move this workload back. I'm not calling it repatriation, I'm calling it rightsizing for the operating model that I have. >> Makes sense. Yeah. >> Go ahead. >> If I missed something, Dave, why we are on this topic of repatriation. I'm actually surprised that we are talking about repatriation as a very big thing. I think repatriation is happening, no doubt, but it's such a small percentage of cloud migration that to me it's a rounding error in my opinion. I think there's a bigger problem. The problem is that people don't know where the cost is. If they knew where the cost was being wasted in the cloud, they could do something about it. But if you don't know, then the easy answer is cloud costs a lot and moving it back to on-premises. I mean, take like Capital One as an example. They got rid of all the data centers. Where are they going to repatriate to? They're all in the cloud at this point. So I think my point is that data observability is one of the places that has seen a lot of traction is because of cost. Data observability, when it first came into existence, it was all about data quality. Then it was all about data pipeline reliability. And now, the number one killer use case is FinOps. >> Maribel, you had a comment? >> Yeah, I'm kind of in violent agreement with both Sanjeev and Keith. So what are we seeing here? So the first thing that we see is that many people wildly overspent in the big public cloud. They had stranded cloud credits, so to speak. The second thing is, some of them still had infrastructure that was useful. So why not use it if you find the right workloads to what Keith was talking about, if they were more static workloads, if it was already there? So there is a balancing that's going on. And then I think fundamentally, from a trend standpoint, these things aren't binary. Everybody, for a while, everything was going to go to the public cloud and then people are like, "Oh, it's kind of expensive." Then they're like, "Oh no, they're going to bring it all on-prem 'cause it's really expensive." And it's like, "Well, that doesn't necessarily get me some of the new features and functionalities I might want for some of my new workloads." So I'm going to put the workloads that have a certain set of characteristics that require cloud in the cloud. And if I have enough capability on-prem and enough IT resources to manage certain things on site, then I'm going to do that there 'cause that's a more cost-effective thing for me to do. It's not binary. That's why we went to hybrid. And then we went to multi just to describe the fact that people added multiple public clouds. And now we're talking about super, right? So I don't look at it as a one-size-fits-all for any of this. >> A a number of practitioners leading up to Supercloud2 have told us that they're solving their cloud complexity by going in monocloud. So they're putting on the blinders. Even though across the organization, there's other groups using other clouds. You're like, "In my group, we use AWS, or my group, we use Azure. And those guys over there, they use Google. We just kind of keep it separate." Are you guys hearing this in your view? Is that risky? Are they missing out on some potential to tap best of breed? What do you guys think about that? >> Everybody thinks they're monocloud. Is anybody really monocloud? It's like a group is monocloud, right? >> Right. >> This genie is out of the bottle. We're not putting the genie back in the bottle. You might think your monocloud and you go like three doors down and figure out the guy or gal is on a fundamentally different cloud, running some analytics workload that you didn't know about. So, to Sanjeev's earlier point, they don't even know where their cloud spend is. So I think the concept of monocloud, how that's actually really realized by practitioners is primary and then secondary sources. So they have a primary cloud that they run most of their stuff on, and that they try to optimize. And we still have forked workloads. Somebody decides, "Okay, this SAP runs really well on this, or these analytics workloads run really well on that cloud." And maybe that's how they parse it. But if you really looked at it, there's very few companies, if you really peaked under the hood and did an analysis that you could find an actual monocloud structure. They just want to pull it back in and make it more manageable. And I respect that. You want to do what you can to try to streamline the complexity of that. >> Yeah, we're- >> Sorry, go ahead, Keith. >> Yeah, we're doing this thing where we review AWS service every day. Just in your inbox, learn about a new AWS service cursory. There's 238 AWS products just on the AWS cloud itself. Some of them are redundant, but you get the idea. So the concept of monocloud, I'm in filing agreement with Maribel on this that, yes, a group might say I want a primary cloud. And that primary cloud may be the AWS. But have you tried the licensed Oracle database on AWS? It is really tempting to license Oracle on Oracle Cloud, Microsoft on Microsoft. And I can't get RDS anywhere but Amazon. So while I'm driven to desire the simplicity, the reality is whether be it M&A, licensing, data sovereignty. I am forced into a multi-cloud management style. But I do agree most people kind of do this one, this primary cloud, secondary cloud. And I guarantee you're going to have a third cloud or a fourth cloud whether you want to or not via shadow IT, latency, technical reasons, et cetera. >> Thank you. Sanjeev, you had a comment? >> Yeah, so I just wanted to mention, as an organization, I'm complete agreement, no organization is monocloud, at least if it's a large organization. Large organizations use all kinds of combinations of cloud providers. But when you talk about a single workload, that's where the program arises. As Keith said, the 238 services in AWS. How in the world am I going to be an expert in AWS, but then say let me bring GCP or Azure into a single workload? And that's where I think we probably will still see monocloud as being predominant because the team has developed its expertise on a particular cloud provider, and they just don't have the time of the day to go learn yet another stack. However, there are some interesting things that are happening. For example, if you look at a multi-cloud example where Oracle and Microsoft Azure have that interconnect, so that's a beautiful thing that they've done because now in the newest iteration, it's literally a few clicks. And then behind the scene, your .NET application and your Oracle database in OCI will be configured, the identities in active directory are federated. And you can just start using a database in one cloud, which is OCI, and an application, your .NET in Azure. So till we see this kind of a solution coming out of the providers, I think it's is unrealistic to expect the end users to be able to figure out multiple clouds. >> Well, I have to share with you. I can't remember if he said this on camera or if it was off camera so I'll hold off. I won't tell you who it is, but this individual was sort of complaining a little bit saying, "With AWS, I can take their best AI tools like SageMaker and I can run them on my Snowflake." He said, "I can't do that in Google. Google forces me to go to BigQuery if I want their excellent AI tools." So he was sort of pushing, kind of tweaking a little bit. Some of the vendor talked that, "Oh yeah, we're so customer-focused." Not to pick on Google, but I mean everybody will say that. And then you say, "If you're so customer-focused, why wouldn't you do a ABC?" So it's going to be interesting to see who leads that integration and how broadly it's applied. But I digress. Keith, at our first supercloud event, that was on August 9th. And it was only a few months after Broadcom announced the VMware acquisition. A lot of people, myself included said, "All right, cuts are coming." Generally, Tanzu is probably going to be under the radar, but it's Supercloud 22 and presumably VMware Explore, the company really... Well, certainly the US touted its Tanzu capabilities. I wasn't at VMware Explore Europe, but I bet you heard similar things. Hawk Tan has been blogging and very vocal about cross-cloud services and multi-cloud, which doesn't happen without Tanzu. So what did you hear, Keith, in Europe? What's your latest thinking on VMware's prospects in cross-cloud services/supercloud? >> So I think our friend and Cube, along host still be even more offended at this statement than he was when I sat in the Cube. This was maybe five years ago. There's no company better suited to help industries or companies, cross-cloud chasm than VMware. That's not a compliment. That's a reality of the industry. This is a very difficult, almost intractable problem. What I heard that VMware Europe were customers serious about this problem, even more so than the US data sovereignty is a real problem in the EU. Try being a company in Switzerland and having the Swiss data solvency issues. And there's no local cloud presence there large enough to accommodate your data needs. They had very serious questions about this. I talked to open source project leaders. Open source project leaders were asking me, why should I use the public cloud to host Kubernetes-based workloads, my projects that are building around Kubernetes, and the CNCF infrastructure? Why should I use AWS, Google, or even Azure to host these projects when that's undifferentiated? I know how to run Kubernetes, so why not run it on-premises? I don't want to deal with the hardware problems. So again, really great questions. And then there was always the specter of the problem, I think, we all had with the acquisition of VMware by Broadcom potentially. 4.5 billion in increased profitability in three years is a unbelievable amount of money when you look at the size of the problem. So a lot of the conversation in Europe was about industry at large. How do we do what regulators are asking us to do in a practical way from a true technology sense? Is VMware cross-cloud great? >> Yeah. So, VMware, obviously, to your point. OpenStack is another way of it. Actually, OpenStack, uptake is still alive and well, especially in those regions where there may not be a public cloud, or there's public policy dictating that. Walmart's using OpenStack. As you know in IT, some things never die. Question for Sanjeev. And it relates to this new breed of data apps. And Bob Muglia and Tristan Handy from DBT Labs who are participating in this program really got us thinking about this. You got data that resides in different clouds, it maybe even on-prem. And the machine polls data from different systems. No humans involved, e-commerce, ERP, et cetera. It creates a plan, outcomes. No human involvement. Today, you're on a CRM system, you're inputting, you're doing forms, you're, you're automating processes. We're talking about a new breed of apps. What are your thoughts on this? Is it real? Is it just way off in the distance? How does machine intelligence fit in? And how does supercloud fit? >> So great point. In fact, the data apps that you're talking about, I call them data products. Data products first came into limelight in the last couple of years when Jamal Duggan started talking about data mesh. I am taking data products out of the data mesh concept because data mesh, whether data mesh happens or not is analogous to data products. Data products, basically, are taking a product management view of bringing data from different sources based on what the consumer needs. We were talking earlier today about maybe it's my vacation rentals, or it may be a retail data product, it may be an investment data product. So it's a pre-packaged extraction of data from different sources. But now I have a product that has a whole lifecycle. I can version it. I have new features that get added. And it's a very business data consumer centric. It uses machine learning. For instance, I may be able to tell whether this data product has stale data. Who is using that data? Based on the usage of the data, I may have a new data products that get allocated. I may even have the ability to take existing data products, mash them up into something that I need. So if I'm going to have that kind of power to create a data product, then having a common substrate underneath, it can be very useful. And that could be supercloud where I am making API calls. I don't care where the ERP, the CRM, the survey data, the pricing engine where they sit. For me, there's a logical abstraction. And then I'm building my data product on top of that. So I see a new breed of data products coming out. To answer your question, how early we are or is this even possible? My prediction is that in 2023, we will start seeing more of data products. And then it'll take maybe two to three years for data products to become mainstream. But it's starting this year. >> A subprime mortgages were a data product, definitely were humans involved. All right, let's talk about some of the supercloud, multi-cloud players and what their future looks like. You can kind of pick your favorites. VMware, Snowflake, Databricks, Red Hat, Cisco, Dell, HP, Hashi, IBM, CloudFlare. There's many others. cohesive rubric. Keith, I wanted to start with CloudFlare because they actually use the term supercloud. and just simplifying what they said. They look at it as taking serverless to the max. You write your code and then you can deploy it in seconds worldwide, of course, across the CloudFlare infrastructure. You don't have to spin up containers, you don't go to provision instances. CloudFlare worries about all that infrastructure. What are your thoughts on CloudFlare this approach and their chances to disrupt the current cloud landscape? >> As Larry Ellison said famously once before, the network is the computer, right? I thought that was Scott McNeley. >> It wasn't Scott McNeley. I knew it was on Oracle Align. >> Oracle owns that now, owns that line. >> By purpose or acquisition. >> They should have just called it cloud. >> Yeah, they should have just called it cloud. >> Easier. >> Get ahead. >> But if you think about the CloudFlare capability, CloudFlare in its own right is becoming a decent sized cloud provider. If you have compute out at the edge, when we talk about edge in the sense of CloudFlare and points of presence, literally across the globe, you have all of this excess computer, what do you do with it? First offering, let's disrupt data in the cloud. We can't start the conversation talking about data. When they say we're going to give you object-oriented or object storage in the cloud without egress charges, that's disruptive. That we can start to think about supercloud capability of having compute EC2 run in AWS, pushing and pulling data from CloudFlare. And now, I've disrupted this roach motel data structure, and that I'm freely giving away bandwidth, basically. Well, the next layer is not that much more difficult. And I think part of CloudFlare's serverless approach or supercloud approaches so that they don't have to commit to a certain type of compute. It is advantageous. It is a feature for me to be able to go to EC2 and pick a memory heavy model, or a compute heavy model, or a network heavy model, CloudFlare is taken away those knobs. and I'm just giving code and allowing that to run. CloudFlare has a massive network. If I can put the code closest using the CloudFlare workers, if I can put that code closest to where the data is at or residing, super compelling observation. The question is, does it scale? I don't get the 238 services. While Server List is great, I have to know what I'm going to build. I don't have a Cognito, or RDS, or all these other services that make AWS, GCP, and Azure appealing from a builder's perspective. So it is a very interesting nascent start. It's great because now they can hide compute. If they don't have the capacity, they can outsource that maybe at a cost to one of the other cloud providers, but kind of hiding the compute behind the surplus architecture is a really unique approach. >> Yeah. And they're dipping their toe in the water. And they've announced an object store and a database platform and more to come. We got to wrap. So I wonder, Sanjeev and Maribel, if you could maybe pick some of your favorites from a competitive standpoint. Sanjeev, I felt like just watching Snowflake, I said, okay, in my opinion, they had the right strategy, which was to run on all the clouds, and then try to create that abstraction layer and data sharing across clouds. Even though, let's face it, most of it might be happening across regions if it's happening, but certainly outside of an individual account. But I felt like just observing them that anybody who's traditional on-prem player moving into the clouds or anybody who's a cloud native, it just makes total sense to write to the various clouds. And to the extent that you can simplify that for users, it seems to be a logical strategy. Maybe as I said before, what multi-cloud should have been. But are there companies that you're watching that you think are ahead in the game , or ones that you think are a good model for the future? >> Yes, Snowflake, definitely. In fact, one of the things we have not touched upon very much, and Keith mentioned a little bit, was data sovereignty. Data residency rules can require that certain data should be written into certain region of a certain cloud. And if my cloud provider can abstract that or my database provider, then that's perfect for me. So right now, I see Snowflake is way ahead of this pack. I would not put MongoDB too far behind. They don't really talk about this thing. They are in a different space, but now they have a lakehouse, and they've got all of these other SQL access and new capabilities that they're announcing. So I think they would be quite good with that. Oracle is always a dark forest. Oracle seems to have revived its Cloud Mojo to some extent. And it's doing some interesting stuff. Databricks is the other one. I have not seen Databricks. They've been very focused on lakehouse, unity, data catalog, and some of those pieces. But they would be the obvious challenger. And if they come into this space of supercloud, then they may bring some open source technologies that others can rely on like Delta Lake as a table format. >> Yeah. One of these infrastructure players, Dell, HPE, Cisco, even IBM. I mean, I would be making my infrastructure as programmable and cloud friendly as possible. That seems like table stakes. But Maribel, any companies that stand out to you that we should be paying attention to? >> Well, we already mentioned a bunch of them, so maybe I'll go a slightly different route. I'm watching two companies pretty closely to see what kind of traction they get in their established companies. One we already talked about, which is VMware. And the thing that's interesting about VMware is they're everywhere. And they also have the benefit of having a foot in both camps. If you want to do it the old way, the way you've always done it with VMware, they got all that going on. If you want to try to do a more cross-cloud, multi-cloud native style thing, they're really trying to build tools for that. So I think they have really good access to buyers. And that's one of the reasons why I'm interested in them to see how they progress. The other thing, I think, could be a sleeping horse oddly enough is Google Cloud. They've spent a lot of work and time on Anthos. They really need to create a certain set of differentiators. Well, it's not necessarily in their best interest to be the best multi-cloud player. If they decide that they want to differentiate on a different layer of the stack, let's say they want to be like the person that is really transformative, they talk about transformation cloud with analytics workloads, then maybe they do spend a good deal of time trying to help people abstract all of the other underlying infrastructure and make sure that they get the sexiest, most meaningful workloads into their cloud. So those are two people that you might not have expected me to go with, but I think it's interesting to see not just on the things that might be considered, either startups or more established independent companies, but how some of the traditional providers are trying to reinvent themselves as well. >> I'm glad you brought that up because if you think about what Google's done with Kubernetes. I mean, would Google even be relevant in the cloud without Kubernetes? I could argue both sides of that. But it was quite a gift to the industry. And there's a motivation there to do something unique and different from maybe the other cloud providers. And I'd throw in Red Hat as well. They're obviously a key player and Kubernetes. And Hashi Corp seems to be becoming the standard for application deployment, and terraform, or cross-clouds, and there are many, many others. I know we're leaving lots out, but we're out of time. Folks, I got to thank you so much for your insights and your participation in Supercloud2. Really appreciate it. >> Thank you. >> Thank you. >> Thank you. >> This is Dave Vellante for John Furrier and the entire Cube community. Keep it right there for more content from Supercloud2.

>>Okay, welcome back everyone. There's the cubes coverage of VMware Explorer, 22 formerly world. We've been here since 2010 and world 2010 to now it's 2022. And it's VMware Explorer. We're here at the CEO, regular writer. Welcome back to the cube. Great to see you in person. >>Yeah. Great to be here in person, >>Dave and I are, are proud to say that we've been to 12 straight years of covering VMware's annual conference. And thank you. We've seen the change in the growth over time and you know, it's kind of, I won't say pinch me moment, but it's more of a moment of there's the VMware that's grown into the cloud after your famous deal with Andy jazzy in 2016, we've been watching what has been a real sea change and VMware since taking that legacy core business and straightening out the cloud strategy in 2016, and then since then an acceleration of, of cloud native, like direction under your leadership at VMware. Now you're the CEO take us through that because this is where we are right now. We are here at the pinnacle of VMware 2.0 or cloud native VMware, as you point out on your keynote, take us through that history real quick. Cuz I think it's important to know that you've been the architect of a lot of this change and it's it's working. >>Yeah, definitely. We are super excited because like I said, it's working, the history is pretty simple. I mean we tried running our own cloud cloud air. We cloud air didn't work so well. Right. And then at that time, customers really gave us strong feedback that the hybrid they wanted was a Amazon together. Right. And so that's what we went back and did and the andjay announcement, et cetera. And then subsequently as we were continue to build it out, I mean, once that happened, we were able to go work with the Satia and Microsoft and others to get the thing built out all over. Then the next question was okay, Hey, that's great for the workloads that are running on vSphere. What's the story for workloads that are gonna be cloud native and benefit a lot from being cloud native. So that's when we went the Tansu route and the Kubernetes route, we did a couple of acquisitions and then we started that started paying off now with the Tansu portfolio. And last but not the least is once customers have this distributed portfolio now, right. Increasingly everything is becoming multi-cloud. How do you manage and connect and secure. So that's what you start seeing that you saw the management announcement, networking and security and everything else is cooking. And you'll see more stuff there. >>Yeah know, we've been talking about super cloud. It's kinda like a multi-cloud on steroids kind a little bit different pivot of it. And we're seeing some use cases. >>No, no, it's, it's a very great, it's a, it's pretty close to what we talk about. >>Awesome. I mean, and we're seeing this kind of alignment in the industry. It's kind of open, but I have to ask you, when did you, you have the moment where you said multicloud is the game changer moment. When did you have, because you guys had hybrid, which is really early as well. When was the Raghu? When did you have the moment where you said, Hey, multicloud is what's happening. That's we're doubling down on that go. >>I mean, if you think about the evolution of the cloud players, right. Microsoft really started picking up around the 2018 timeframe. I mean, I'm talking about Azure, right? >>In a big way. >>Yeah. In a big way. Right. When that happened and then Google got really serious, it became pretty clear that this was gonna be looking more like the old database market than it looked like a single player cloud market. Right. Equally sticky, but very strong players all with lots of IP creation capability. So that's when we said, okay, from a supplier side, this is gonna become multi. And from a customer side that has always been their desire. Right. Which is, Hey, I don't want to get locked into anybody. I want to do multiple things. And the cloud vendors also started leveraging that OnPrem. Microsoft said, Hey, if you're a windows customer, your licensing is gonna be better off if you go to Azure. Right. Oracle did the same thing. So it just became very clear. >>I am, I have gone make you laugh. I always go back to the software mainframe because I, I think you were here. Right. I mean, you're, you're almost 20 years in. Yeah. And I, the reason I appreciate that is because, well, that's technically very challenging. How do you make virtualization overhead virtually non-existent how do you run any workload? Yeah. How do you recover from, I mean, that's was not trivial. Yeah. Okay. So what's the technical, you know, analog today, the real technical challenge. When you think about cross cloud services. >>Yeah. I mean, I think it's different for each of these layers, right? So as I was alluding to for management, I mean, you can go each one of them by themselves, there is one way of Mo doing multi-cloud, which is multiple clouds. Right. You could say, look, I'm gonna build a great product for AWS. And then I'm gonna build a great product for Azure. I'm gonna build a great product for Google. That's not what aria is. Aria is a true multi-cloud, which means it pulls data in from multiple places. Right? So there are two or three, there are three things that aria has done. That's I think is super interesting. One is they're not trying to take all the data and bring it in. They're trying to federate the data sources. And secondly, they're doing it in real time and they're able to construct this graph of a customer's cloud resources. >>Right. So to keep the graph constructed and pulling data, federating data, I think that's a very interesting concept. The second thing that, like I said is it's a real time because in the cloud, a container might come and go like that. Like that is a second technical challenge. The third it's not as much a technical challenge, but I really like what they have done for the interface they've used GraphQL. Right? So it's not about if you remember in the old world, people talk about single pan or glass, et cetera. No, this is nothing to do with pan or glass. This is a data model. That's a graph and a query language that's suited for that. So you can literally think of whatever you wanna write. You can write and express it in GraphQL and pull all sorts of management applications. You can say, Hey, I can look at cost. I can look at metrics. I can look at whatever it is. It's not five different types of applications. It's one, that's what I think had to do it at scale is the other problem. And, and >>The, the technical enable there is just it's good software. It's a protocol. It's >>No, no, it's, it's, it's it's software. It's a data model. And it's the Federation architecture that they've got, which is open. Right. You can pull in data from Datadog, just as well as from >>Pretty >>Much anything data from VR op we don't care. Right? >>Yeah. Yeah. So rego, I have to ask you, I'm glad you like the Supercloud cuz you know, we, we think multi-cloud still early, but coming fast. I mean, everyone has multiple clouds, but spanning this idea of spanning across has interesting sequences. Do you data, do you do computer both and a lot of good things happening. Kubernetes been containers, all that good stuff. Okay. How do you see the first rev of multi-cloud evolving? Like is it what happens? What's the sequence, what's the order of operations for a client standpoint? Customer standpoint of, of multicloud or Supercloud because we think we're seeing it as a refactoring of something like snowflake, they're a data base, they're a data warehouse on the cloud. They, they say data cloud they'd they like they'll tell us no, you, we're not a data. We're not a data warehouse. We're data cloud. Okay. You're a data warehouse refactored for the CapEx from Amazon and cooler, newer things. Yeah, yeah, yeah. That's a behavior change. Yeah. But it's still a data warehouse. Yeah. How do you see this multi-cloud environment? Refactoring? Is there something that you see that might be different? That's the same if you know what I'm saying? Like what's what, what's the ne the new thing that's happening with multi-cloud, that's different than just saying I'm I'm doing SAS on the cloud. >>Yeah. So I would say, I would point to a, a couple of things that are different. Firstly, my, the answer depends on which category you are in. Like the category that snowflake is in is very different than Kubernetes or >>Something or Mongo DB, right? >>Yeah. Or Mongo DB. So, so it is not appropriate to talk about one multi-cloud approach across data and compute and so, so on and so forth. So I'll talk about the spaces that we play. Right. So step one, for most customers is two application architectures, right? The cloud native architecture and an enterprise native architecture and tying that together either through data or through networks or through et cetera. So that's where most of the customers are. Right. And then I would say step two is to bring these things together in a more, in a closer fashion and that's where we are going. And that is why you saw the cloud universal announcement and that's already, you've seen the Tansu announcement, et cetera. So it's really, the step one was two distinct clouds. That is just two separate islands. >>So the other thing that we did, that's really what my, the other thing that I'd like to get to your reaction on, cause this is great. You're like a masterclass in the cube here. Yeah, totally is. We see customers becoming super clouds because they're getting the benefit of, of VMware, AWS. And so if I'm like a media company or insurance company, if I have scale, if I continue to invest in, in cloud native development, I do all these things. I'm gonna have a da data scale advantage, possibly agile, which means I can build apps and functionality very quick for customers. I might become my own cloud within the vertical. Exactly. And so I could then service other people in the insurance vertical if I'm the insurance company with my technology and create a separate power curve that never existed before. Cause the CapEx is off the table, it's operating expense. Yep. That runs into the income statement. Yep. This is a fundamental business model shift and an advantage of this kind of scenario. >>And that's why I don't think snowflakes, >>What's your reaction to that? Cuz that's something that, that is not really, talk's highly nuanced and situational. But if Goldman Sachs builds the biggest cloud on the planet for financial service for their own benefit, why wouldn't they >>Exactly. >>And they're >>Gonna build it. They sort of hinted at it that when they were up on stage on AWS, right. That is just their first big step. I'm pretty sure over time they would be using other clouds. Think >>They already are on >>Prem. Yeah. On prem. Exactly. They're using VMware technology there. Right? I mean think about it, AWS. I don't know how many billions of dollars they're spending on AWS R and D Microsoft is doing the same thing. Google's doing the same thing we are doing. Not as much as them that you're doing oral chair. Yeah. If you are a CIO, you would be insane not to take advantage of all of this IP that's getting created and say, look, I'm just gonna bet on one. Doesn't make any sense. Right. So that's what you're seeing. And then >>I think >>The really smart companies, like you talked about would say, look, I will do something for my industry that uses these underlying clouds as the substrate, but encapsulates my IP and my operating model that I then offer to other >>Partners. Yeah. And their incentive for differentiation is scale. Yeah. And capability. And that's a super cloud. That's a, or would be say it environment. >>Yeah. But this is why this, >>It seems like the same >>Game, but >>This, I mean, I think it environment is different than >>Well, I mean it advantage to help the business, the old day service, you >>Said snowflake guys out the marketing guys. So you, >>You said snowflake data warehouse. See, I don't think it's in data warehouse. It's not, that's like saying, you >>Know, I, over >>VMware is a virtualization company or service now is a help desk tool. I, this is the change. Yes. That's occurring. Yes. And that you're enabling. So take the Goldman Sachs example. They're gonna run OnPrem. They're gonna use your infrastructure to do selfer. They're gonna build on AWS CapEx. They're gonna go across clouds and they're gonna need some multi-cloud services. And that's your opportunity. >>Exactly. That's that's really, when you, in the keynote, I talked about cloud universal. Right? So think of a future where we can go to a customer and say, Mr. Customer buy thousand scores, a hundred thousand cores, whatever capacity you can use it, any which way you want on any application platform. Right. And it could be OnPrem. It could be in the cloud, in the cloud of their choice in multiple clouds. And this thing can be fungible and they can tie it to the right services. If they like SageMaker they could tie it to Sage or Aurora. They could tie it to Aurora, cetera, et cetera. So I think that's really the foundation that we are setting. Well, I think, I >>Mean, you're building a cloud across clouds. I mean, that's the way I look at it. And, and that's why it's, to me, the, the DPU announcement, the project Monterey coming to fruition is so important. Yeah. Because if you don't have that, if you're not on that new Silicon curve yep. You're gonna be left behind. Oh, >>Absolutely. It allows us to build things that you would not otherwise be able to do, >>Not to pat ourselves on the back Ragu. But we, in what, 2013 day we said, feel >>Free. >>We, we said with Lou Tucker when OpenStack was crashing. Yeah. Yeah. And then Kubernetes was just a paper. We said, this could be the interoperability layer. Yeah. You got it. And you could have inter clouding cuz there was no clouding. I was gonna riff on inter networking. But if you remember inter networking during the OSI model, TCP and IP were hardened after the physical data link layer was taken care of. So that enabled an entire new industry that was open, open interconnect. Right. So we were saying inter clouding. So what you're kind of getting at with cross cloud is you're kind of creating this routing model if you will. Not necessarily routing, but like connection inter clouding, we called it. I think it's kinda a terrible name. >>What you said about Kubernetes is super critical. It is turning out to be the infrastructure API so long. It has been an infrastructure API for a certain cluster. Right. But if you think about what we said about VSE eight with VSE eight Kubernetes becomes the data center API. Now we sort of glossed over the point of the keynote, but you could do operations storage, anything that you can do on vSphere, you can do using a Kubernetes API. Yeah. And of course you can do all the containers in the Kubernetes clusters and et cetera, is what you could always do. Now you could do that on a VMware environment. OnPrem, you could do that on EKS. Now Kubernetes has become the standard programming model for infrastructure across. It >>Was the great equalizer. Yeah. You, we used to say Amazon turned the data center through an API. It turns, turns of like a lot of APIs and a lot of complexity. Right. And Kubernetes changed. >>Well, the role, the role of defacto standards played a lot into the T C P I P revolution before it became a standard standard. What the question Raghu, as you look at, we had submit on earlier, we had tutorial on as well. What's the disruptive enabler from a defacto. What in your mind, what should, because Kubernetes became kind of defacto, even though it was in the CNCF and in an open source open, it wasn't really standard standard. There's no like standards, body, but what de facto thing has to happen in your mind's eye around making inter clouding or connecting clouds in a, in a way that's gonna create extensibility and growth. What do you see as a de facto thing that the industry should rally around? Obviously Kubernetes is one, is there something else that you see that's important for in an open way that the industry can discuss and, and get behind? >>Yeah. I mean, there are things like identity, right? Which are pretty critical. There is connectivity and networking. So these are all things that the industry can rally around. Right. And that goes along with any modern application infrastructure. So I would say those are the building blocks that need to happen on the data side. Of course there are so many choices as well. So >>How about, you know, security? I think about, you know, when after stuck net, the, the whole industry said, Hey, we have to do a better job of collaborating. And then when you said identity, it just sort of struck me. But then a lot of people tried to sort of monetize private reporting and things like that. So you do you see a movement within the technology industry to do a better job of collaborating to, to solve the acute, you know, security problems? >>Yeah. I think the customer pressure and government pressure right. Causes that way. Yeah. Even now, even in our current universe, you see, there is a lot of behind the scenes collaboration amongst the security teams of all of the tech companies that is not widely seen or known. Right. For example, my CISO knows the AWS CSO or the Microsoft CSO and they all talk and they share the right information about vulnerability attacks and so on and so forth. So there's already a certain amount of collaboration that's happening and that'll only increase. Do, >>Do you, you know, I was somewhat surprised. I didn't hear more in your face about security would, is that just because you had such a strong multi-cloud message that you wanted to get, get across, cuz your security story is very strong and deep. When you get into the DPU side of things, the, you know, the separation of resources and the encryption and I'll end to end >>I'm well, we have a phenomenal security story. Yeah. Yeah. Tell security story and yes. I mean I'll need guilty to the fact that in the keynote you have yeah, yeah, sure time. But what we are doing with NSX and you will hear about some NSX projects as you, if you have time to go to some of the, the sessions. Yeah. There's one called project, not star. Another is called project Watchman or watch, I think it's called, we're all dealing with this. That is gonna strengthen the security story even more. Yeah. >>We think security and data is gonna be a big part of it. Right. As CEO, I have to ask you now that you're the CEO, first of all, I'd love to talk about product with you cuz you're yeah. Yeah. We just great conversation. We want to kind of read thet leaves and ask pointed questions cuz we're putting the puzzle together in real time here with the audience. But as CEO, now you have a lot of discussions around the business. You, the Broadcom thing happening, you got the rename here, you got multi-cloud all good stuff happening. Dave and I were chatting before we came on this morning around the marketplace, around financial valuations and EBIDA numbers. When you have so much strategic Goodwill and investment in the oven right now with the, with the investments in cloud native multi-year investments on a trajectory, you got economies of scale there. >>It's just now coming out to be harvest and more behind it. Yeah. As you come into the Broadcom and or the new world wave that's coming, how do you talk about that value? Cuz you can't really put a number on it yet because there's no customers on it. I mean some customers, but you can't probably some for form. It's not like sales numbers. Yeah. Yeah. How do you make the argument to the PE type folks out there? Like EBIDA and then all the strategic value. What's the, what's the conversation like if you can share any, I know it's obviously public company, all the things going down, but like how do you talk about strategic value to numbers folks? >>Yeah. I mean, we are not talking to PE guys at all. Right. I mean the only conversation we have is helping Broadcom with >>Yeah. But, but number people who are looking at the number, EBIDA kind of, >>Yeah. I mean, you'd be surprised if, for, for example, even with Broadcom, they look at the business holistically as what are the prospects of this business becoming a franchise that is durable and could drive a lot of value. Right. So that's how they look at it holistically. It's not a number driven. >>They do. They look at that. >>Yeah. Yeah, absolutely. So I think it's a misperception to say, Hey, it's a numbers driven conversation. It's a business driven conversation where, I mean, and Hawk's been public about it. He says, look, I look at businesses. Can they be leaders in their market? Yeah. Because leaders get, as we all know a disproportionate share of the economic value, is it a durable franchise that's gonna last 10 years or more, right. Obviously with technology changes in between, but 10 years or more >>Or 10, you got your internal, VMware talent customers and >>Partners. Yeah. Significant competitive advantage. So that's, that's really where the conversation starts and the numbers fall out of it. Got it. >>Okay. So I think >>There's a track record too. >>That culture >>That VMware has, you've always had an engineering culture. That's turned, you know, ideas and problems into products that, that have been very successful. >>Well, they had different engineering cultures. They're chips. You guys are software. Right. You guys know >>Software. Yeah. Mean they've been very successful with Broadcom, the standalone networking company since they took it over. Right. I mean, it's, there's a lot of amazing innovation going on there. >>Yeah. Not, not that I'm smiling. I want to kind of poke at this question question. I'll see if I get an answer out of you, when you talk to Hawk tan, does he feel like he bought a lot more than he thought or does he, did he, does he know it's all here? So >>The last two months, I mean, they've been going through a very deliberate process of digging into each business and certainly feels like he got a phenomenal asset base. Yeah. He said that to me even today after the keynote, right. Is the amazing amount of product capability that he's seeing in every one of our businesses. And that's been the constant frame. >>But congratulations on that. >>I've heard, I've heard Hawk talk about the shift to, to Mer merchant Silicon. Yeah. From custom Silicon. But I wanted to ask you when you look at things like AWS nitro yeah. And graviton and train and the advantage that AWS has with custom Silicon, you see Google and Microsoft sort of Alibaba following suit. Would it benefit you to have custom Silicon for, for DPU? I mean, I guess you, you know, to have a tighter integration or do you feel like with the relationships that you have that doesn't buy you anything? >>Yeah. I mean we have pretty strong relationships with in fact fantastic relationships with the Invidia and Intel and AMD >>Benon and AMD now. >>Yeah. Yeah. I mean, we've been working with the Pendo team in their previous incarnations for years. Right, right. When they were at Cisco and then same thing with the, we know the Melanox team as well as the invi original teams and Intel is the collaboration right. From the get go of the company. So we don't feel a need for any of that. We think, I mean, it's clear for those cloud folks, right. They're going towards a vertical integration model and select portions of their stack, like you talked about, but there is always a room for horizontal integration model. Right. And that's what we are a part of. Right. So there'll be a number of DPU pro vendors. There'll be a number of CPU vendors. There'll be a number of other storage, et cetera, et cetera. And we think that is goodness in an alternative model compared to a vertically integr >>And yeah. What this trade offs, right. It's not one or the other, I mean I used to tell, talk to Al Shugar about this all the time. Right. I mean, if vertically integrated, there may be some cost advantages, but then you've got flexibility advantages. If you're using, you know, what the industry is building. Right. And those are the tradeoffs, so yeah. Yeah. >>Greg, what are you excited about right now? You got a lot going on obviously great event. Branding's good. Love the graphics. I was kind of nervous about the name changed. I likem world, but you know, that's, I'm kind of like it >>Doesn't readily roll off your phone. Yeah. >>I know. We, I had everyone miscue this morning already and said VMware Explorer. So >>You pay Laura fine. Yeah. >>Now, I >>Mean a quarter >>Curse jar, whatever I did wrong. I don't believe it. Only small mistake that's because the thing wasn't on. Okay. Anyway, what's on your plate. What's your, what's some of the milestones. Do you share for your employees, your customers and your partners out there that are watching that might wanna know what's next in the whole Broadcom VMware situation. Is there a timeline? Can you talk publicly about what? To what people can expect? >>Yeah, no, we, we talk all the time in the company about that. Right? Because even if there is no news, you need to talk about what is where we are. Right. Because this is such a big transaction and employees need to know where we are at every minute of the day. Right? Yeah. So, so we definitely talk about that. We definitely talk about that with customers too. And where we are is that the, all the processes are on track, right? There is a regulatory track going on. And like I alluded to a few minutes ago, Broadcom is doing what they call the discovery phase of the integration planning, where they learn about the business. And then once that is done, they'll figure out what the operating model is. What Broadcom is said publicly is that the acquisition will close in their fiscal 23, which starts in November of this year, runs through October of next year. >>So >>Anywhere window, okay. As to where it is in that window. >>All right, Raghu, thank you so much for taking valuable time out of your conference time here for the queue. I really appreciate Dave and I both appreciate your friendship. Congratulations on the success as CEO, cuz we've been following your trials and tribulations and endeavors for many years and it's been great to chat with you. >>Yeah. Yeah. It's been great to chat with you, not just today, but yeah. Over a period of time and you guys do great work with this, so >>Yeah. And you guys making, making all the right calls at VMware. All right. More coverage. I'm shot. Dave ante cube coverage day one of three days of world war cup here in Moscone west, the cube coverage of VMware Explorer, 22 be right back.