>>From New York city. It's the cube covering. Welcome to the new edge brought to you by systems. >>Hey, welcome back everybody. Jeff, Rick here with the cube. We are in Manhattan at the top of Goldman Sachs. It is a great view if you ever get an opportunity to come up here, I think 43 floors over the Hudson you could see forever. But this is the cloud events. So the clouds are here and we're excited to be here is the Penn Penn Sandow launch in the name of the event is welcome to the new edge, which is a pretty interesting play. We hear a lot about edge but we haven't really heard of that company really focusing on the edge as their primary go to market activity and really thinking about the edge first. So we're excited to have the cofounder cube Olam and many time guests a Sony Gian Deni. She's the co founder and chief business officer. So many great to see you. Good to see you too. >>And our hosts here at Goldman Sachs is uh, Josh Matthews. He's a managing director of technology at Goldman. Josh. Great to see you. You too. And thank you and thanks for hosting us. Nice. A nice place to come to work every day. So great conversation today. Congratulations on the launch of the company over two years in stealth mode. Talk a little bit about that. What is it like to be in stealth mode for so long and you guys raised big money, you've got a big team, you're doing heavy duty technology. What's it been like to finally open up the curtains and tell everybody what you've been? >>It's clearly very interesting and exciting. Normally it's taken me nine months to deliver a baby this time it's been two and a half years of being instilled while we have been getting ready for this baby to come out. So it's phenomenally exciting that too to be sharing the stage with our customers and our investors and our strategic partners. >>Yeah, I thought it was pretty interesting that you're launching with customers and when you really told the story on stage of how early you engaged with Josh and his team, um, first I want to get your kinda your perspective. Why were you doing that so early and what did that ultimately do with some of the design decisions that you guys made? And then we'll come back to Josh as to, you know, his participation. >>So I think whenever you conduct technology transitions, having a sense from customers that have the ability to look out two to three years is very important because when you're capturing market transitions, doing it with customer inputs is far more relevant than going about it alone. Uh, the other key thing about this architectural shift is that it allows the flexibility for every customer to go take pieces of how they want to bring the cloud architectures and bring it into their environment. So understanding that use case and understanding the compelling reasons of what problems both technological and business can be solving and having that perspective into the product definition and the design and the influence that customers like Josh you've had is why we are sitting here and talking about them in production. Uh, as opposed to, yeah, we're thinking about where we are. We are looking at it from a proof of concept perspective. Right. >>And Josh, your, your perspective, you said earlier today that, you know, as long as a sign is involved, you're, you're, uh, you're happy to jump in and see what she's been working on. So how, >>you know, how did you get involved, how did they reach out to you and, and what is it like working on, you know, technology so early in its development that you get to actually have some serious influence? Well, it's an amazing opportunity, um, to get exactly what you want, um, exactly what you know is going to solve problems for the business here. Um, you know, and the other thing is, you know, we've worked with this team, uh, through almost every spinning. Uh, I think it was a little young for the, maybe the first one. Um, but, uh, otherwise this team has worked with them through at least 15 years or more. So we knew the track record for execution and then for us on this product, I mean, it was an opportunity because it's truly a startup. Um, you know, Sony and the team brought us in. >>Uh, we kind of just put out problems on the table that we were trying to solve and then, you know, they came up with the product and the idea and we were able to put together, you know, yeah, these are our priority one, two, three that we want to go for. And you know, we've just been developing alongside them. So both software and, you know, driving what the feature set is. Right. So what were some of those problems guys? Price seemed like forever ago when you started this conversation, but as you kind of looked forward a couple of years back that you could see that were coming, that you needed addressed. You know, it's funny, we started with kind of like, well we think containerization is going to be explosive and, and you know, really everything's on virtual machines or bare metal, mostly virtual machines. So one, you know, as containers come out, how do we track them, secure them, um, how do we even secure, uh, you know, the virtual machines and our environment cause they're, you know, over almost a quarter million of them. >>The idea of being able to put, um, network policy, that's I would say incorruptible, not actually on the server, but at, you know, that's why we use firewalls, right? So solving that security problem was number one. The other one was being able to have the telemetry to see what's happening, what's changing, um, and troubleshoot at, you know, at the network layer from every single server. Again, it's all about scale. Like things were just scaling and the throughput's going up, traditional methods of being able to see what's on your network. You can't look in the middle, it just can't keep up. It's just speeds and feeds. So being able to push those things to the edge. And then lastly, it really happened more, um, through the process here. But about a year and a half ago, um, we began segmenting our network the same way a 5g provider does with a technology called segment routing. >>And we just said, that's kind of our follow on technologies to, you know, put the network in the server and put this segment routing capability all the way out at the edge. So, you know, some things we foresaw and other things we've just developed. You know, it's been, it's been two and a half years. So, um, it's been a great partnership and you know, I think more, more features will come. Well Sony, you and the team, but it's been talked about all day long, have have a history of multiple times that you've kind of brought these big transformational technologies. Um, head what, what did you guys see a couple of years back and kind of this progression, you saw this opportunity >>to do something a little bit different than you've done in the past, which is actually go out, raise, raise around and uh, and do a real startup. What was the opportunity that you saw this? >>So we saw a number of challenges and opportunities. At the same time, we, we clearly saw that, uh, the cloud architectures that have been built by the leaders, like the incumbents like AWS today have a lot of the intelligence that is being pushed into their, their respective compute platforms. Uh, and we also noticed that at the same time, while that was what was needed to build the first generation of the cloud, the new age applications, and even as gardener has predicted that 75% of all enterprise data and applications will be processed at the edge by 2025. If that happens, then you need that intelligence at the edge. You need the ability to go do it where the action is, which is at the edge. And very consistently we found that the architectures, including scale out storage, we're also driving the need for this intelligence to be on in a scale-out manner. >>So if you're going to scale out computing, you need the services to be going hand in hand with that scale. Our computer architecture for the enterprises so they can simplify their architectures and bring the cloud models that have only existed in the cloud world, into their own data centers and their own private clouds. So there were these technology transitions we saw were coming down the pike. It's easier said now in 2019 it wasn't so simple in 2017 because we had to look at these multiple technology transitions. And surprisingly, when we call those things out, as we were shaping the company's strategy, getting validation of the use cases from customers like Josh was pivotally important because it was for the validating that this would be the direction that the enterprises and the cloud customers would be taking. So the reason you start with a vision, you start with looking at where the technology transitions are going to be occurring and getting the customers that are looking farther out validated plays a very important role so that you can go and focus on the biggest problems that you need to go and solve. Right, right. >>It just seems like the, the, the big problem, um, for most layman's is, is the old one, which, why networking exists in the first place, which is do you bring the data to the compute or do you bring the compute to the data? And now as you said, in kind of this hyper distributed world, um, that's not really a viable answer either one, right? Because the two are blended and have to be together so that you don't necessarily have to move one to the other or the other back the other direction. So, and then the second piece that you talked about over and over in your, in your presentation with security and you know, everybody talks about security all the time. Everybody gets hacked every day. Um, and there's this constant theme that security has to be baked in, you know, kind of throughout the process as opposed to kind of bolted on at the end. You guys took that approach from day, just speak >>it into the architecture. Yes. That was crucially important because when you are trying to address the needs of the enterprise, particularly in regulated markets like financial services, you want to be in a position where you have thought about it and baked it into the platform ground up. Uh, and so when we are building the program of a process, so we had the opportunity to go put the right elements on it. In order to make it tamper proof, we had to go think about encrypting all the traffic and communication between our policy manager and the distributed services platforms at the edge. We also then took it a step further to say, now if there were to be a bad actor that were to attack from an operating system vulnerability perspective, how do we ensure that we can contain that bad actor as opposed to being propagated over the infrastructure? So those elements are things you cannot bolt on at design time, or when you need to go put those into the design day one, right. Only on top of that foundation, then can you build a very secure set of services, whether it's encryption, whether it's distributed via services, so on and so forth. >>Uh, and Josh, I'm curious on your take as we've seen kind of software defined everything, uh, slowly take over as opposed to, you know, kind of single purpose machines or single purpose appliances, et cetera. Yep. Really a different opportunity for you to control. Um, but also to see a lot of talk today about, about policy management. A lot of talk about, um, observability and as you said now even segmentation of the networks, like you segment the nodes and you segment everything else. You know, how, how do you see this kind of software defined everything continuing to evolve and what does it enable you to do that you can't do with just a static device? I mean, the approach we took, um, we started like, you know, years ago, about six years ago was saying we can get computers, uh, deployed for our applications. No problem. Uh, and you know, at, at on demand and in our internal cloud, now we can do it as a hybrid cloud solution. >>One of the biggest problems we had in software defined was how do you put security policy, firewall policy, um, with that compute and in, you know, our industry, there's lots of segmentation for material nonpublic information. Um, compliance, you know, it could be internet facing, B2B facing. Uh, we do that today. We program various firewall vendors automatically. Uh, we allow our application developers to create, um, these policies and push them through as code and then program the firewall. What we were really looking to do here is distribute that. So we F day one in getting pen Sandow into production was to use our uh, our firewall system. It's called pinnacle. We, um, we programmed from pinnacle directly into the Penn Songdo Venice manager via API and then it, you know, uses its inventory systems to push those things out. So for us, software defined has been around, I like to call it the store front, but for the developer it's network policy, it's load balancing. >>Um, and, and that's really what they see. Those are the big products on the net. Everything else is just packet forwarding to them. So we wanted with pen Sandow at least starting with security to have that bar set day one and then get, you know, all the benefits of scale, throughput and having the policies close to the, on the edge. You know, we're back to talking about the edge. We want to right there with the, with the deployment, with the workload or the application. And that's, that's what we're doing right off the bat. Yeah. What are the things you mentioned in your talk was w is, you know, kind of in the theme of atomic computing, right? You want to get smaller and smaller units so that you can apply and redeploy based on wherever the workload is and in the change. And you said you've now been able to, you know, basically take things out of dedicated, you know, kind of a dedicated space, dedicated line and dedicated job so that you can now put them in a more virtualized situation. >>Exactly. Grab more resources as you need them. Well, you'd think the architecture, I mean even just theater of the mind is just, you're saying, I'm going to put this specific thing that I have to secure behind these firewalls. So it's one cabinet of computers or a hundred it's still behind a set of firewalls. It's a very North, South, you know, get in and get out here. You're talking about having that same level of security and I think that's novel, right? There hasn't been, if you look at virtual firewalls or you know, IP tables on Linux, I mean it's corruptible. It's, it's, it can be attacked on the computer. And once it's, you know, once you've been attacked in that, that that attack vector has been, you know, hit your, your compromised. This is a separate management plane. Um, you know, separate control plane. The server doesn't see it. >>That security is provided. It's at scale, it's East, West. The more computers that have the pen Sandow, you know, architecture inside of them, the, you know, the wider you can go, right. And then the North South goes away. I'm just curious to get your perspective. Um, as you know, everyone is a technology company. At the same time, technology budgets are going down, people are hard to hire. Uh, your data is growing exponentially and everything's a security threat. Yes. So as you get up in the morning, get ready to drive to work and you're drinking your coffee, I mean, how do you, you know, kind of communicate to make sure to senior management knows kind of what your objectives are in this, this kind of ongoing challenge to do more with less. And it, even though it's an increasingly strategic place or is it actually is what the company does now, it just happens to wrap it around your plane services or financial services or travel or whatever. >>Uh, I think your eye, and I had said it to John before, um, it has to come from that budget has to come from somewhere. So I think a combination of, of one that's less, well, I'll say the one that's easier to quantify is you're going to take budget from say appliance manufacturer and move it to a distributed edge and you're going to hopefully save some money while you do it. Um, you're going to do it at scale. You're gonna do it at, you know, high throughput and the security is the same or better. So that's, that's one, that's one place to take capital from. The other one is to say, can I use the next computer? Yes. Because I don't have to deploy these other new computers behind this stack of firewalls. Is there agility there? Is there efficiency, um, on my buying less servers and using, you know, more of what I have and doing it, you know, able to deploy faster. >>And it's harder to quantify. I think if you could, you know, over time, see I bought 20% less server, uh, capacity or, you know, x86 capacity, that's a savings. And the other one that's very hard to quantify, but it's always nice to have the development community. And we've had it recently where they say, Hey, this took me a month to deploy instead of a year. Um, and you know, the purchase cycles, uh, you know, for procurement and deployment, they're long, you know, in enterprise you want them to be quick, but they're really not. So all of those things add up. And that's the story. You know, I would tell, you know, any manager, right? Yeah, >>yeah. I think, you know, the old historic way that utilization rates were just so, so, so, so low between CPU and memory, everything else. Cause if nothing else, because to get another box, you know, could take a long time. Yeah. Well, final, final question for you, Tony. You talked about architectures and being locked into architectures and you and you talked about you guys are already looking forward, you know, to kind of your next rev, your next release, kind of your next step forwards. What, where do you see kind of the direction, don't give away any secrets, but um, you know, kind of where you guys going. What are your priorities now that you've launched? You got a little bit more money in the bank. >>Well, our biggest priorities will be to focus on customer success is to make sure that the customer journey is indeed replicable at scale, is to enable the partner's success. Uh, so in addition to Goldman Sachs, the ability to go and replicate it across the federated markets, whether it's global financial services, healthcare, federal, and partnering with each B enterprise so that they can on their platform, amplify the value of this architecture, not just on the compute platforms but on, in other areas. And the third one clearly is for our cloud customers is to make sure that they are in a position to build a world class cloud architecture on top of which then they can build their own, deliver their own services, their own secret sauces, uh, so that they can Excel at whatever that cloud is. Whether it's to become the leading edge platform as a service customer, whether it is to be the leading edge of software's a service platform customer. So it's all about the execution as a, as you heard in that room. And that's fundamentally what we're going to strive to be, is to be a great execution machine and keep our heads down and focused on making our customers and our partners very successful. >>Well, certainly, congratulations again to you and the team on the launch today. And Josh, thank you for hosting this terrific event and being an early customer. Yeah. Yeah. Happy to be. Alright. I'm Jetta. Sone. Josh, we're the topic. Goldman Sachs at the Penn Sandow the new welcome to the new edge. Thanks for watching. We'll see you next time.