(upbeat music) >> Welcome back to Supercloud 22. I'm John Furrier, your host of "The Cube." We're here for a live performance in studio bringing all the thought leaders around this concept of Supercloud, which is a consortium of the smartest people in the industry, the the Cloudaratti some say, or just people in the field building out next generation Cloud technologies for businesses, for the industry, you know, software meets infrastructure at scale and platforms. All great stuff. We have an expert here, Cube alumni and friend of ours, Howie Xu, VP of machine learning and AI at Zscaler, hugely successful company, platform, whatever you want to call it. They're definitely super clouding in their own. Howie, great to see you. Thanks for spending time with us to unpack and grock the direction of the industry that we see. We call it Supercloud. >> Hey, John, great to be back. I'm expecting a nice very educational and interesting conversation here again. >> Yeah, well, you know, one of the things I love talking with you about is you're deep on the technology side, as well as you got the historian view like we do. We've seen the movies before, we've seen the patterns, and now we're seeing structural change that has happened, that's cloud. Thank you very much, AWS. And as your GCP and others. Now we're seeing structural change happening in real time and we want to talk about it as it's happening. This is the purpose of this event. And that is that Cloud is one. Okay, great Cloud operations, on premises and Edge are emerging. Software is open source. It's the perfect storm for innovation and new things are emerging. You're seeing companies like Snowflake, and Databricks, and Zscaler all building great products. But now it's not one thing anymore. It's a lot of things going on. So what is your take on Supercloud? How do you see this evolving? What is some of the structural change that's happening in your mind? >> Yeah, so when you first reached out a few weeks ago about this event, I was like, "Hey, what is Supercloud." I know you tweeted a little bit here and there, but I never really, you know, double clicked, right. So I actually listened to some of your episodes you know, the previous conversations. You know, I would say the way you define Supercloud is it's not just the multi-cloud. The multi-Cloud is probably one aspect of it, right. You know, it's actually more beyond that, right. You know, a little bit, you know, towards past, a little bit more towards the flexibility, and then, you know, including, and also you want to include the on-prem, the edge, not just the Big 3 cloud, right. So there is a lot of the, let's say hybrid, more inclusive, right. So, the way I look at it is it's now very different from my imagination of where the Cloud would be, should be 10, 12 years ago. Because, you know, at that time it was, you know, on-prem dominant and then we say, hey, let's go cloud. I never for a second thought, you know, we would've ditch the on-prem completely, right. You know, on-prem has its own value. It's own kind of characteristics we wanted to keep, right. But the way we went for the last 10 years is, hey, Cloud, Cloud everywhere. We embrace Cloud. You know, the way I look at architecture is always very much like a pendulum, right? We swung from decentralized in the mainframe days, you know back in the days, to more distributed, right, PC, kind of a architecture, you know, servers in your own data center. And then to the now, the Cloud, the Big 3 Cloud in particular, right. I think in the next 10, 15, 20 years, it will swing back to more decentralized, more distributed architecture again. Every time you have a swing, because there is some fundamental reason behind that, we all knew the reason behind the current swing to the Cloud. It's because hey, the on-prem data center was too complex, right. You know, too expensive, right. You know, it would've take at least the six months to get any business application going, right. So compared to Cloud, a swipe a credit card, frictionless, you know, pay as you go, it's so great. But I think we are going to see more and more reason for people to say, "Hey, I need a architecture the other way around because of the decentralized the use case," right. Web3 is one example. Even though Web3 is still, you know, emerging right, very, very early days. But that could be one reason, right? You mentioned the Zscaler is kind of a Supercloud of its own, right? We always embrace public Cloud, but a lot of the workloads is actually on our own, you know, within our own data center. We take advantage of the elasticity of the public Cloud, right. But we also get a value, get a performance of our private Cloud. So I want to say a company like Zscaler taking advantage of the Supercloud already, but there will be more and more use cases taking advantage. >> And the use cases are key. Let me just go back and share something we had on the panel earlier in the day, the Cloudaratti Panel. Back in 2008, a bunch of us were getting together and we kind of were riffing, oh yeah, the future's going to be web services and Clouds will talk to each other, workloads can work across this (indistinct) abstraction layer, APIs is going to be talking to each other. A little bit early but we tried to think about it in terms of the preferred architecture. Okay, way too early. Yeah. AWS was just getting going, really kind of pumping on all cylinders there, getting that trajectory up. But it was use case driven. The nirvana never happened. I mean, we were talking Supercloud back then with the Cloudaratti group and we were thinking, okay, hey this is cool. But it was just an evolutionary thing. So I want to get your reaction. Today, the use cases are different. It's not just developers deploying on public Cloud to get all those greatness and goodness of the Cloud, to your point about Zscaler and others, there's on premises use cases and edge use cases emerging. 5g is right there. That's going to explode. So, the use cases now are all Cloud based. Again, this is an input into what we're seeing around Supercloud. How do you see that? What's your reaction to that? And how do you see that evolving so that the methodologies and all the taxonomies are in place for the right solution? >> Right, I mean, you know, some of the use cases are already here, you know, have been here for the last few years. And again, I mentioned a Zscaler, right. The reason that a Zscaler needs the on-prem version of it is because it's impossible to route all the traffic to the Big 3 Cloud, because they're still far away. Sometimes you need the presence much closer to you in order for you to get the level of the performance latency you want, right. So that's why Zscaler has, you know, so many data center of our own instead of leveraging the public Cloud, you know, for most part. However, public Cloud is still super important for Zscaler. I can tell you a story, right. You know, two years ago, you know, at the beginning of the pandemics, everyone started working from home suddenly, right. You are talking about Fortune 500 companies with 200,000 employees, suddenly having 200,000 employees working from home. Their VPN architecture is not going to support that kind of the workload, right? Even Zscaler's own architecture or the presence is not enough. So overnight we just, having so many new workloads, to support this work from home, the zero trust network for our customers, literally overnight. So it wouldn't have happened without public Cloud. So we took advantage of the public Cloud. Yet at the same time, for many, many use cases that Zscaler is paying attention to in terms of the zero trust architecture, the latency, the latency guarantee aspect, the cost is so important. So we kind of take it advantage of both. >> Yeah, definitely. >> Today you may say, hey, you know, Zscaler is one of the, not a majority of the companies in terms of the Cloud adoption or public Cloud adoption, right. But I can say that, yeah, that's because it's more infrastructure, security infrastructure. It's a little bit different for some of the communication applications, right. Why not just support everything on the public Cloud? That's doable today. However, moving forward next to 5, 10, 15 years, we expect to see Web3 kind of the use cases to grow more and more. In those kind of decentralized use cases, I can totally see that we, you know, the on-prim presence is very important. >> Yeah. One of the things we're seeing with Supercloud that we're kind of seeing clarity on is that there's a lot of seamless execution around, less friction around areas that require a PhD or hard work. And you're seeing specialty Superclouds, apps, identity data security. You're also seeing vertical clouds, Goldman Sachs doing financial applications. I'm sure there'll be some insurance. People in these verse. Building on top of the CapEx on one Cloud really fast and moving to others. So that's clearly a trend. The interesting thing I want to get your thoughts on, Howie, on an architectural basis is in Cloud, public Cloud generally, SaaS depends on IAS. So there's an interplay between SaaS and the infrastructures of service and pass as well. But SaaS and IAS, they solve a lot of the problems. You mentioned latency. How do you see the interplay of these Superclouds that utilize the SaaS IS relationship to solve technical problems? So in architecturally, that's been a tight integration on these Clouds, but now as you get more complexity with Supercloud, how do you see SaaS applications changing? >> Yeah, I view the Supercloud is actually reduced the complexity. The reason I'm saying that is, think about it in the world where you have predominantly public Cloud kind of the architecture, right? 10 years ago, AWS has probably 20 services. Now they probably have, you know, more than 1,000 services. Same thing with Azure, same thing with GCP. I mean, who can make sense out of it, right. You know, if you just consume the eyes or the Big 3 Cloud service as is. You know, you need a PhD these days to make sense all of them. So the way I think about Supercloud or where, you know it is going, is it has to provide more simplicity, a better way for people to make sense out of it, right. Cause if I'm an architecture and I have to think, hey, this is a public Cloud, this is a multi-Cloud, and by the way, certain things need to be run on the on-prem. And how do I deal with the uniform nature of it? My mind would blow up. So I need a higher level abstraction. That higher level abstraction will hide the complexity of the where it is, which vendor. It will only tell me the service level, right. You know, we always say, you know, the Cloud is like electricity. I only wanted to know is that like 110 volt or 220, 240, whatever that is. I don't really want to know more than that, right. So I want to say a key requirement for the Supercloud is it's reduced the complexity, higher level abstraction. It has to be like that. >> And operational consistencies at the bottom. Howie, we have one minute left. I want to get your thoughts. I'd like you to share what you're working on that you're excited about. It doesn't have to be with Zscaler. As you see the Supercloud trend emerging, this is the next generation Cloud, Cloud 2.0, whatever we want to call it, it's happening. It's changing. It's getting better. What are you excited about? What do you see as really key inflection point variables in this big wave? >> Yeah. One of the things I really like, what I heard from you in the past about Supercloud is a Supercloud is not just a one Cloud or one vendor. It's almost like every company should have its own Supercloud, right. You're talking about JP Morgan, Goldman Sachs of the world, that they need to have their own Supercloud. Zscaler and their security vendors, they may have their own Cloud. So I think every Fortune 500, Fortune 2,000 companies will have its own Supercloud. So I'm excited about that. So why that's important? We also say that, you know, in the next 10, 20 years, AI machine learning is going to help us a lot, right. So without Supercloud, it's very hard to do AI machine learning. 'Cause if you don't have a place that you know where the data is, and then it's pretty hard. And in the context of Supercloud, I totally foresee that the AI model will follow the data. If the data is in the cloud, it will go there. If the data is on-prem, it will go there. And then the Supercloud will hide the complexity of it. So if you ask me, my passion is leveraging AI machine learning to change the world, but Supercloud will make that easier, right. If you think about why Google, Facebook of the world, are able to leverage AI better than 99% of the rest of the world, because they figure out the Supercloud for themselves, right. And I think now it's the time for the rest of the Fortune 500, of Fortune 2,000 company to figure out its own Supercloud strategy. What is my Supercloud? I need to have my own Supercloud. Each company needs to have its own Supercloud. That's how I see it. >> Howie, always great to have you on. Thanks so much for spending the time and weighing in on this really important topic. We're going to be opening this up. It's not over. We're going to continue to watch the change as it unfolds and get an open community perspective. Thank you so much for being a great expert in our network and community. We really appreciate your time. >> Thank you for having me. >> Okay. Okay, that's it. We'll be up with more coverage here, Supercloud event, after this short break. I'm John Furrier, host of "The Cube." Thanks for watching. (upbeat music)