 Welcome back to SuperCloud 22. I'm John Furrier, host of theCUBE. We're here at Palo Alto for a big event, SuperCloud 22. We've got a great ecosystem conversation here. Ramesh Prabhagaran, who's the co-founder and CEO of Promethio. Ramesh, great to see you. Thanks for coming on. Thanks for having me, John. So I wanted to bring you in because we had previous CUBE conversations around cloud networking latency. You also have some pedigree, Viptella, the folks in the industry know that it's been a deep tech company. You've been around the block. You've seen the movie before. You've seen the tech trends. You've seen the hype. You've seen the fluff. Where's the meat on the bone with SuperCloud in your opinion? So it starts with what enterprises are struggling with, right? And if you take a very simple example, it's actually quite fresh in my mind because I was just having this conversation this morning. A large bank has an application sitting in AWS, right? And they have to provide the application access to the treasury, to their suppliers, to ticker feeds, to all their downstream partners and so on and so forth. Guess what? They don't control where all those things are. They're in very different regions, very different clouds. And so whether you like it or not, you have a problem here, right? And so it starts with, for the particular bank, what are the capabilities that they need, right? And so AWS provides a whole host of native capabilities, but they still need to build a few more things on top. So going by essentially the definition of SuperCloud, even within a single cloud, you need to build a few more capabilities on top. That gets worsened by the fact that now you need to provide access to various other clouds, various other regions and so forth. So whether we like it or not, this movie is here to stay. What's the difference between SuperCloud and Multicloud? Because Multicloud, I've been saying, is not necessarily a market yet. Correct, yes. So SuperCloud is essentially the cloud native capabilities provided by the hyperscalers get you probably 30, 40% of the way, right? But then in order to deliver on a care about, right? In our case, from cloud networking standpoint, that is experience, that's performance, reliability, zero trust access and so forth, you have to take that a little bit further. And so we have vendors like us that actually build capabilities on top of the hyperscalers, right? Now, even if you think of a single cloud, how you build that is different on AWS than its own Azure than on GCP. But do the customers care? No, they want to be able to consume it in exactly the same way across all of them. So whether it's multi-cloud or a single cloud, you have a problem that is white space on top of the hyperscalers capabilities that you just need to build. And what problems is that solving today? Because again, Multicloud, I've not seen the problem. I kind of get what's happening. Multiple clouds do exist. Use cases matter, maybe best to breed, but they're standalone. They're not really interoperating, so to speak. So people have been successful on public cloud. That's correct, yes. For use cases. Absolutely, so even if you take a single cloud for example, right? You have multiple problems to address. So let's take the example of I have users coming from various different regions around the globe and I have apps that I've spread maybe not across like all clouds, but single cloud maybe multiple regions, right? Now I have a reach problem, which is I need to go from where the user is to where the application is sitting. I have an experience problem because if my spinning wheel shows up, I'm going to go crazy. I have a security problem because I want to make sure it's only me that have access to it, right? But does the cloud provider solve for this entirely? No, they give you the nuts, the bowls, what we call as essentially, what you need is a latte. They give you really nice coffee beans, not just one flavor, 20 flavors of those. They give you raw sugar and a few other things. They give you five different flavors of milk, but you got to make your own latte. So that's what you do. And this is where the infrastructure transformations happen. Exactly. And the super pass layers, Dave Vellante and I have talked about in cloud is you have to integrate a native cloud, which is beautiful, it's integrated, everything works together. There's a lot of lattes to be made or expressoes. I mean, tons of great things there. So big check marks, double check, gold star for AWS. Correct. All good. Now on premises, we found that hybrid is a steady state. Exactly. That's cloud operation. Now you got the edge. Where does the super cloud strategy come in? For the folks watching there, it's like, hey, okay, I get that, but I don't want to just buy into another vendor's hype. Absolutely. I got to build my own cloud to your point about the lattes. Correct. Build their own infrastructure and application environment to power the developers. Exactly. And hybrid is here to stay as you pointed out, right, John? So I have my data center and let's say when most folks start out, they start to like a single region of a single cloud, right? And what are you most concerned about there? Hey, can I migrate? Can I start to build applications in the public cloud? And all you care about is can I talk back into my data center? Like as long as some basic hygiene is there, that's all they care about, right? The problem happens when you go from kind of the first five EC2 instances to 50 to 100, then you have a few other things that you need to care about, right? That's really kind of where the super cloud capabilities start to come in, right? Because you have the cloud native things, you can make that work for the first few days, but then after that, you need augmented capabilities. So Ramesh, some people will say, hey John, super cloud, okay, it's funny, ha ha ha. But isn't it just SAS? No, SAS is a delivery mechanism, right? And so there is the capability and that is how do you want to consume, right? And so capabilities are cloud native, capabilities are a piece of software, capabilities are Kubernetes cluster form factor and so forth. How do you want to consume? Maybe it's a package form factor, it is a SAS, it could even be PAS if it's sitting in the element and so forth, right? And so you really want to distinguish those two. And that's how we see the industry evolve. Can super clouds be specialty clouds? Like a snowflake of super cloud, is Goldman Sachs financial cloud a super cloud? Absolutely, right? So super cloud is not like a conglomeration of multiple capabilities, right? It can be for a specific use case, it can be for a specific functionality. So we consider our capabilities by the definition as a super cloud in networking, right? In cloud networking and prosimo. So does that solve the entirety of what I want to do in the cloud? No, absolutely not. There's data, there's compute, there's a whole bunch of other things, but for a specialty, you do have some super cloud. In fact, I had a note here, I was going to ask you, when will there be specialty clouds, apps, identity data, security, networking? We will see those. Absolutely, yeah, and those are slowly starting to brew, right? So you have identity as one, you have networking as one, you have the zero trust piece as another one, you have data as another one. So when all these things come together, absolutely, that's what the enterprise customers care about. So I love infrastructure as code that drove a lot of the evolution and revolution of DevOps. When are we going to see security as code and network as code? Or is it there? No, network as code for sure, it's already there. It's probably in its early innings, I would say, but we are starting to see that already. The reason for that is really simple. Enough CIOs have yelled at their networking teams to say, my app guys can get this done three, four times a day, you get this done once a week, right? And so that has actually driven quite a bit of innovation, it's slow, right? And so that's driven quite a bit of innovation. It all starts with, hey, can I build a Terraform provider? And then just integrate into Terraform, but it doesn't stop there, right? There's a whole bunch of additional capabilities at the end, troubleshooting, a whole bunch of things that need to come together. But I would say network as code has already started to take shape. Miss, that's a great point about specialty clouds. What about vertical clouds too? Cause you got insurance, oil and gas, thin tech, both sides of the stack can have specialty clouds. Absolutely, yeah. So what's driving specialty clouds, right? Some of it is compliance, mainly because you just have to shard the data and when you shard the data, the entirety gets sharded, right? Some of it driven by use case, because some are a little more serverless service mesh and intelligence focus, some are a little more infrastructure focused. So you do see that taking off. I would say we've seen a whole lot more on the horizontal side, less on the vertical side, but that's really happening, right? I think that to me indicates a super cloud. The fact that the diversity of the application on the clouds themselves, someone could be spending, say Liberty Mutual or Goldman Sachs, they were once spending that as capex. Exactly. Now it's OPEX, so they become a service provider. So if you have scale with data and expertise, you become a super cloud by default. Exactly. And you don't have to pay for the capex, you're already paying in. Exactly, yeah. And that's what Snowflake basically did with data warehouse. That's right, yeah. I mean, they're basically a data warehouse. Exactly. Refactored on the cloud and then go, whoa, let's go to Azure. Yeah, and where does that data decide that you ask that question? No, right? You just assume that, hey, with respect to if it's a single cloud in multiple regions, it's there. If it's stretched to multiple clouds, yes, it's just there, but you talk about like that. In our Cloud Erotic panel earlier, we talked about how companies are going fast on one native cloud because they don't want to have multiple development teams and different ops teams. They go all in, say, mostly AWS wins this, unless especially as your productivity software or SQL database. Go hard in on Amazon, get speed and velocity, get that flywheel win, get scale, get value, then go to Azure, provide that same value to that marketplace and other clouds. Then the next dot to connect is, can the customer have the same experience across the clouds? That's where it gets interesting. Exactly. What's your thoughts on that? Actually, it gets interesting even when they go from a single cloud in a single region to multiple regions, right? And the more spread out the regions are, you have requirements around application performance, application experience and so forth. So suddenly the networking conversation starts to become an experience in a performance conversation. The security conversation starts to become a zero trust conversation and so forth. And so you do see that interesting shift that's happening that makes it exactly. And then that gets worsened by the fact that now you have multiple clouds, multiple regions and then. So you get regions, clouds. And then you have edge locations now. You mentioned edge. This is why I think multi cloud is BS because this is all coming so fast. You got to get your super cloud first. Exactly. Then you extend into what it looks like a multi vendor or multi faceted environment that should be automated by that time. Exactly. So it's evolutionary. We're not there yet. Exactly. So you agree no market yet. That's right, yes. So unless it's like the super large enterprises where we have seen a really good mix of multiple different clouds, the super large enterprises where each business unit is free to choose the cloud of their choice for the application developers because they just like a certain cloud, right? And there you find yourself. For negotiations. For negotiations, right? And so, exactly. So there you find yourself in a healthy mix. It's not like you're 80, 10, 10. It's a healthy mix of three different clouds, right? But vast majority of the enterprises, they have a concerted strategy. I have a primary cloud because that's where two big CEOs shake hands and sign multi billion dollar deals, right? It's just a song with Howie Shew who's now a Zscaler, former VM where I probably know Howie. He's a legend in the community as well. We were talking about the old days of the data center. And you remember that we'll go back to our, into our historical views of experience. Back when the data center became popular, versus the glass house. Mainframes, the mini computers, it became a complex environment. You had to have pretty much a PhD or serious networking or some sort of technical background. And then IT was born, the local area networks, the mini computers and the PCs, change that dynamic. IT was born, okay? And let's just say it. Most IT guys aren't PhDs. Exactly. So what's happened there is democratization and the operation side of that wave. We're kind of going through it now, don't you think with cloud? Like you got to be super smart to wrangle the data. I mean, some of the data pipelining stuff is super complex and there's snowflake. Yeah, absolutely. And your data bricks. And largely depends on the maturity, right? Like, so once you pass a certain scale in the cloud, the caribos start to be very different. The caribos are, how can I operate this at scale? Because I might have started off with a relatively inefficient infrastructure, right? But now if I start to operate that at scale with like thousands of VPCs and so forth, somebody is looking at an AWS build there and going, oh, no, no, no, we're not going to do that. We're getting to the good part now. So here's where I wanted to get to because we're kind of getting there. The proof points of SuperCloud is IT like operations, easy, not overstaffed and maybe an SRE model, one to many. Yeah, exactly. What are the proof points do you see that would be evidence that SuperCloud is working? So in a well-functional model where we have seen enterprises take the applications that they care about and then move that into the public cloud or build it organically. If they have staffed their team, I think a good leading indicator is if they have staffed the team so that there are a bunch of guys who understand what it means for cloud native capabilities. There are a bunch of guys who then put it together and then you look at the care about, right? Ultimately, at the end of the day, the goal, if you go higher up in the layers, is it about application experience? Is it about kind of reducing the blast radius of my security? Is it about my data cleanliness and hygiene? You don't care about kind of how the pipelining works underneath the covers or how do I put a transit gateway and this and that together? No, that's not what you care about. Kind of the outcomes and... Paul Moritz, you said at VMware when he was there, you say the hard and tough, no one talks about what's in an Intel processor. Exactly. I mean, it just works. Exactly, yeah. And it's what applications you build on top of that Intel processor that actually makes it more powerful, right? And so the first evidence I would say is kind of how is the team structured? The second evidence would be kind of what are the care about for the guys that are building these applications, right? Because even the application developers more than the application, they care about kind of is it helping? Is it delivering on the experience? Is it being used the way it's supposed to? Is it value? Exactly, right? And those are not areas that the cloud providers are solely focused on, right? Like you don't see an AWS or an Azure dashboard show that particular thing for the entirety of the application. They'll tell you for the ATR services that you use, here's the SLA for each one of these services. And that's where the customer has to build it. Exactly, right? Now, does that give you the full picture? No, it doesn't. Somebody has to pull this together. Somebody has to aggregate this together and then make sense as to whether this is working or not, right? So whether you call it super cloud, whether you call it kind of the caribots on top of the cloud native stuff, they're all the same. I'm glad you guys came up with the name for this and I think it's going to be here to stay. Well, thank you for sharing your expertise. It's got a great background in this area. And I think you guys are right on the front wave of this new change. I think a little bit early, but that's good. But don't be too early. Yeah, exactly. No, and that's really important, right, John? So you don't want to be too early. Certainly don't want to be too late. But at the same time, the pace at which things are evolving are fast enough that you will see. I think if we have this conversation even three months from now, it might be a very different conversation. People want to go fast and they don't want to get stuck with a vendor. They made a bad choice that slows them down because they got problems to solve things to build. Ramesh, thanks for coming on SuperCloud 22. We're breaking it all down. We're exposing it out to everyone. We're discussing it. We're going to challenge it. But ultimately it is a thing. SuperCloud 22, thanks for watching. Wonderful, thanks John.