 Welcome back to SuperCloud 2, this is Dave Vellante. We're here exploring the intersection of data and analytics in the future of cloud. And in this segment, we're going to look at the evolution of cloud and try to test some of the SuperCloud concepts and assumptions with Brian Grace Lee as the founder and co-host along with Aaron Delp of the popular Cloudcast program. Amazing series, if you're not already familiar with it. The Cloudcast is one of the best ways to keep up with so many things going on in our industry. Enterprise tech, platform engineering, business models, obviously cloud developer trends. Crypto, Web 3.0, sorry, Brian. I know that's a sore spot, but Brian, thanks for coming on the program, really appreciate it. Yeah, great to be with you, Dave. Happy New Year, and great to be back with everybody with SiliconANGLE again this year. Yeah, we love having you on. We miss working with you day to day, but I want to start with Grace Lee's theorem, which basically says, I'm a paraphrase. For the most part, nothing new gets introduced in the enterprise tech business. Patterns repeat themselves, maybe get applied in new ways. And you know this industry well, when something comes out that's new, if you take virtualization, for example, I've been around forever with mainframes, but then VMware applied it to solve a real problem in the client service system. And then it's like, okay, this is awesome. We get really excited. Then after a while, we push the architecture, we break things, introduce new things to fix the things that are broken and start adding new features. Oftentimes you do that to acquisition. So has the cloud become that sort of thing in a super cloud sort of same wine new bottle following Grace Lee's theorem? Yeah, I think there's some of both of it. I hate to be the sort of it depends sort of answer, but I think to a certain extent, obviously cloud in and of itself was kind of revolutionary in that, it wasn't that you couldn't rent things in the past. It was just being able to do it at scale, being able to do it with such amazing self-service. And then kind of the proliferation of like, look at how many services I can get from one cloud, whether it was Amazon or Azure or Google. And then we slip back into the things that we know, we go, oh, well, okay, now I can get computing on demand, but now it's just computing or I can get database on demand and it's got some of the same limitations of database. It's still, I have to think about IOPS and I have to think about caching and other stuff. So I think we do go through that. And then we have these sort of next paradigms that come along. So serverless was another one of those where it was like, okay, it seems sort of new. I don't have to, again, it was another level of like, I don't have to think about anything. And I was able to do that because there was either, greater bandwidth available to me or compute got cheaper. And what's been interesting is not the sort of, that specific thing, serverless in and of itself is just another way of doing compute. But the fact that it now gets applied as sort of a no ops model to, again, like how do I provision a database? How do I think about, do I have to think about the location of a service? Does that just get taken care of for me? So I think the super cloud concept and I did a thing and you and I have talked about it behind the scenes that maybe a better name is super app for something like Snowflake or other. But I think we're seeing these sort of evolutions over and over again of what were the big bottlenecks? How do we solve those bottlenecks? And I think the big thing here is it's never, it's very rarely that you can take the old paradigm of what the thing was, the concept was, and apply it to the new model. So I'll just give you an example. So something like VMware, which we all know wildly popular, wildly used. But when we apply like a super cloud concept to VMware, the concept of VMware has always been around a cluster, right? It's some finite number of servers, you sort of manage it as a cluster. And when you apply that to the cloud and you say, okay, there's, for example, VMware in the cloud, it's still the same concept of a cluster of VMware. But yet when you look at some of these other services that it would fit more into the super cloud kind of paradigm, whether it's a Snowflake or a MongoDB Atlas or maybe what Cloudflare is doing at the edge, those things get rid of some of those old paradigms. And I think that's where stuff you start to go, oh, okay, this is very different than before. Yes, it's still computing or storage or data access, but there's a whole nother level of something that we didn't carry forward from the previous days. And that really kind of breaks the paradigm. And so that's the way I think I started to think about, are these things really brand new? Yes and no, but I think it's when you can see that big, that thing that you didn't leave behind isn't there anymore, you start to get some really interesting new innovation come out of it. Yeah, and that's why, lift and shift is okay, but when you talk to practitioners, they'll say, I really didn't change my operating model. And so I just kind of moved it into the cloud. There were a couple of some benefits, but it was maybe one zero, not three zeros that I was looking for. You know, we always talk about what's great about cloud and the agility and all the other wonderful stuff that we know, what's not working in cloud. You know, tie it into multi-cloud, you know, in terms of, you know, you hear people talk about multi-cloud by accident. Okay, and that's true, what's not great about cloud. And then I want to get into, you know, is multi-cloud really a problem or is it just sort of vendor hype? But what's not working in cloud? I mean, you mentioned serverless, and serverless is kind of narrow, right? For a lot of stateless things, but what's not great about cloud? Well, I think there's a few things that if you ask most people, they don't love about cloud. I think, you know, we can argue whether or not sort of this consolidation around a few cloud providers has been a good thing or a bad thing. I think, regardless of that, you know, we are seeing, we are hearing more and more people that say, look, you know, the experience I used to have with cloud when I went to, for example, an Amazon and there was a, you know, a dozen services, it was easy to figure out what was going on. It was easy to figure out what my billing looked like. You know, now they've become so widespread, the number of services they have, you know, the number of stories you just hear are people who went, oh, I started a service over in US West and I can't find it anymore. It's on a different screen and I know I just got billed for it. Like, so I think the sprawl of some of the clouds has gotten, has created a user experience that a lot of people are frustrated with. I think that's one thing. We, you know, we see people like DigitalOcean and we see others who are saying, hey, we're gonna be that simplified version. So there's always that yin and yang. I think people are super frustrated at network costs, right? So, you know, and that's kind of a lot of it, the center of maybe why we do or don't see more of these super cloud services is just, you know, in the data center, as an application owner, I didn't have to think about, well, where does this go to? Where are my users? Yes, somebody took care of it, but when those things become front and center, that's super frustrating. That's the one area that we've seen absolutely no cost savings, cost reduction. So I think that frustrates people a lot. And then I think the third piece is just, you know, we're, we saw, we went from super centralized IT organizations, which, you know, for decades was how it worked. It was part of the reason why the cloud expanded and became a thing, right? Sort of shadow IT and I can't get things done. And then now what we've seen is sort of this proliferation of little pockets of groups that are, that are your IT for lack of a better thing, whether they're called platform engineering or SRE or DevOps, but we have this proliferate expansion, explosion, if you will, of groups that if I'm an app dev team, I go, hey, you help me make this stuff run, but then the team next to you has another group and they have another group. And so you see this explosion of, you know, we don't have any standards in the company anymore. And so sort of self-service has, has created its own nightmare to a certain extent for a lot of larger companies. Yeah, thank you for that. So, you know, I want to, I want to explore this multi-cloud, you know, by accident thing. And is it real problem? You hear that a lot from vendors. And we've been talking about super cloud as this unifying layer across cloud. Yeah. You know, but when you talk to customers, a lot of them are saying, yes, we have multiple clouds in our organization, but my group, we have mono cloud. We know the security, you know, edicts. We know how to, you know, deal with the primitives, whether it's, you know, S3 or Azure blob or whatever it is. And we're very comfortable with this. It's, that's how we're simplifying. So do you, do you think this is really a problem? Does it have merit that we need that unifying layer across clouds? Or is it just too early for that? I think, yeah, I think what you, what you've laid out is, is basically how the world has played out. People have picked a cloud for a specific application or a series of applications. Yeah. I think if you talk to most companies, they would tell you, you know, holistically, yes, we're multi cloud, not maybe not necessarily on, I don't necessarily love the phrase where people say like, well, it happened by accident. I think it happened on purpose, but we got to multi cloud, not in the way that maybe that vendors, you know, perceived, you know, kind of laid out a map for it. So it was, it was, well, you will, you will lay out this sort of super cloud framework. We didn't call it that back then. We just called it sort of multi cloud. Maybe it was Kubernetes or maybe it was whatever. And different groups, because central IT kind of got disbanded or got fragmented, it turned into go pick the best cloud for your application for what you need to do for the business. And then, you know, multiple years later, it was like, oh, hold on, I've got 20% in Google and 50% in AWS. And I've got 30% in Azure. And, you know, it's, yeah, it's been evolution. I don't know that it's, I don't know it's a mistake. I think it's now groups trying to figure out, like, should I make sense of it? You know, should I try and standardize and I backward standardize into stuff? I think that's going to be a hard thing for companies to do. Cause I think they feel okay with where the applications are. They just happen to be in multiple clouds. I want to run on something by you and you guys, you and Aaron have talked about this thing. You know, still depending on who, what's keynote you listen to, small percentage of the workloads are actually in cloud. And when you were with us at Wikibon, I think we called it true private cloud. And we looked at things like Nutanix and there were a lot of other examples of companies that were trying to replicate the hyperscale experience on-prem. And we would evaluate that, you know, beyond virtualization. And so we sort of defined that. But I think what's, maybe what's more interesting than super cloud across clouds is if you include that on-prem estate because that's where most of the work is being done. That's where a lot of the proprietary tools have been built, a lot of data, a lot of software. So maybe there's this concept of sending that true private cloud to true hybrid cloud. So I actually think hybrid cloud in some cases is the more interesting use case for so-called super cloud. What are your thoughts on that? Yeah, I think there's a couple aspects to it. I think if we were to go back five or six years even maybe even a little further and look at what a data center looked like even if it was just, hey, we're a data center that runs primarily on VMware and we use some of their automation versus what you can, even what you can do in your data center today, the gains that people have seen through new types of automation through Kubernetes, through get ops and a number of these things, like they've gotten significantly further along in terms of I can provision stuff really well, I can do multi-tenancy, I can do self-service. Is it still hard? Yeah, because those things are hard to do but there's been significant progress there. I still look for kind of that killer application that sort of lighthouse use case of hybrid applications between data center and between cloud. I think we see some stuff where backup is a part of it. So you use the cloud for storage, maybe use the cloud for certain kinds of resiliency, especially on maybe front end load balancing and stuff. But I think what we get into is this being hung up on hybrid cloud or multi-cloud as a term and go like, look, what are you trying to measure? Are you trying to measure efficiency of IT usage? Are you trying to measure how quickly can I give these business, these application teams that are part of a line of business resources that they need? I think if we start measuring in that way, we would look at, you'd go, wow, that used to be weeks and months. Now we got rid of these boards that after review everything, every time I want to do a change management type of thing, we've seen a lot more self-service. I think those are the things we want to measure on. And then to your point of, where do these super cloud applications fit in? I think there are a bunch of instances where you go, look, I have a global application. I have a thing that has to span multiple regions. That's where the super cloud concept really comes into play. We used to do it in the data center, right? We'd had all sorts of technologies to help with that. I think you can now start to do it in the cloud. You know, one of the other things I'm trying to understand, I'd love to get your thoughts on this. Do you think, you again, have talked about this, I'm with you, it's like, how is it that Google's losing $3 billion a year or whatever? Because when you go back and look at Amazon, when they were at that level of revenue, where Google is today, they were making money, you know, and they're actually growing faster, by the way. So it's kind of interesting what's happening with Google. But the reason I bring that up is, trying to understand if you think the hyperscalers will ever be motivated to create standards across clouds, and that may be a play for Google. I mean, obviously with Kubernetes, it was like a Hail Mary and kind of made them relevant. Where would Google be without Kubernetes? But then did it achieve the objectives? We could have that conversation some other time. But do you think the hyperscalers will actually say, okay, we're going to lean in and create these standards across clouds, because customers would love that, I would think, but it would sub-optimize their competitive advantage. What are your thoughts? I think, you know, on the surface, I would say they probably aren't. I think if you asked them the question, they would say, well, first and foremost, you know, we do deliver standards. So we deliver a standard SQL interface or a standard Kubernetes API or whatever. So in that perspective, you know, we're not locking you into an Amazon-specific database or a Google-specific database, you can argue about that. But, and I think to a certain extent, like they've been very good about, hey, we're going to adopt the standards that people want a lot of times, the open source standards. I think the problem is, let's say they did come up with a standard for it. I think you still have the problem of the costs of migration and the longer you've, I think their bet is basically the longer you've been in some cloud, and again, the more data you sort of compile there, the data gravity concept, there's just going to be a natural thing that says, okay, the hurdle to get over to say, look, we want to move this to another cloud, become so cost prohibitive that they don't really have to worry about, you know, oh, I'm going to get into a war of standards and so forth. I think they sort of realize like that's the, that's the flywheel that the cloud creates. And, you know, unless they want to get into a road where they just cut bandwidth costs, like it just kind of won't happen. You know, I think we've even seen, and you know, the one example I'll use, and I forget the name of it off the top of my head, but there's a Google service. I think it's like BigQuery external or something along those lines that allows you to say, look, you can use BigQuery against like S3 buckets and against other stuff. And so I think the cloud routers have kind of figured out, I'm never going to get the application out of that other guy's cloud or the other, you know, the other cloud, but maybe I'm going to have to figure out some interesting ways to sort of work with it. And, you know, it's a little bit, it's a little janky, but that might be, you know, a moderate step that gets sort of gets customers where they want to be. Yeah, or, you know, be interested if you ever see AWS, for example, running its database in other clouds, you started even Oracle is doing that with Azure, which is a form of super cloud. My last question for you was, I want you to get thinking about sort of how the future plays out, you know, to think about some of the companies that we've put forth as super cloud. And by the way, this has been a criticism of the concept, Charleston, everything is super cloud, which if true would defeat the purpose, of course. And so with the community effort, we really tried to put some guardrails down on the essential characteristics, the deployment models, you know, so for example, running across multiple clouds with the purpose build pass, creating a common experience, metadata intelligence that solves a specific problem. I mean, the example I often use is Snowflake's governed data sharing, but yeah, Snowflake, Databricks, Cloudflare, Cohesity, you know, I just mentioned Oracle and Azure, these and others, they certainly claim to have that common experience across clouds. But my question is, again, I come back to it, do customers need this capability? You know, is MonoCloud the way to solve that problem? What's your, what are your thoughts on how this plays out in the future of I guess, Paz, apps and cloud? Yeah, I think a couple of things. So from a technology perspective, I think, you know, the companies you name, the services you've named have sort of proven that the concept is viable and it's viable at a reasonable size, right? These aren't completely niche businesses, right? They're multi-billion dollar businesses. So I think there's a subset of applications that, you know, maybe a bigger than a niche set of applications are gonna use these types of things. A lot of what you talked about is very data centric and that's fine. That layer is figuring that out. I think we'll see messaging types of services. So like Derek Collison, Derek Collison, Cindiya company runs a sort of a super cloud for messaging applications. So I think there'll be places where it makes a ton of sense. I think the thing that I'm not sure about and because again, we've been now 10 plus years of sort of super low, you know, low and not low interest rates in terms of being able to do things is a lot of these things come out of research that have been done previously, then they get turned in and maybe somewhat of an open source project and then they can become something. You know, will we see as much investment into the next Snowflake if, you know, that the interest rates are three or four times they used to be, do we see VCs doing it? So that's the part that worries me a little bit is I think we've seen what's possible. I think, you know, we've seen companies like what those services are. I think I read yesterday, Snowflake was saying like their biggest customers are growing at 50 or 60% like the value they get out of it is becoming exponential. And it's just a matter of like, will the economics allow the next big thing to happen? Cause some of these things are pretty cost, you know, expensive to get started. So I'm bullish on the idea. I don't know that it becomes, I think it's okay that it's still sort of, you know, niche plus plus in terms of the size of it because, you know, if we think about all of IT, it's still, you know, even microservices is a small part of, you know, bigger things. But I'm still really bullish on the idea. I like that it's been proven. I'm a little wary, like a lot of people with the economics of, you know, what might slow things down a little bit. But yeah, I think the future is going to involve super cloud somewhere, whenever people end up calling it. And you know what I'm just gonna say. But I don't think it goes away. I don't think it's a fad. I think it is something that people see tremendous value and it's just, it's got to be, you know, for what you're trying to do, your application-specific thing. You're making a great point on the funding of innovation and we're entering a new era of public policy as well. R&D tax credit is now shifting to, you know, you're going to have to capitalize that over five years now. And that's something that goes back to the 1950s and many people would argue that's at least in part what has helped the United States be so, you know, competitive in tech. But Brian, always great to talk to you. Thanks so much for participating in the program. Great to see you. Thanks, Dave. Appreciate it. Good luck with the rest of the show. Thank you. All right, this is Dave Vellante for John Furrier in the entire Cube community. Stay tuned for more content from SuperCloud 2.