 Hello everyone, welcome to another great CUBE conversation here in Palo Alto. Our CUBE studios, I'm John Furrier, we've got a great conversation with a friend, Sar Goli, CUBE alumni. Sar is an executive coaching startups, investing, lives in Silicon Valley, sees a lot of the landscape, knows networking, knows cloud. Sar, great to see you. Thanks for coming. Thanks for being here. It's great to be here again, as always. You know, love having you come in and be an analyst for us and help us squint through some of the big mega trends that we're following. As you know, we believe that cloud is here to stay and I'm always chirpin' certainly online about how Amazon clearly way ahead of Microsoft. Google has a great network and you and I are going to talk more about that. But the game is just getting started and compute is phase one of the game of cloud. Completely been commoditized, yet serverless. A lot of great things are helping developers. This next wave of networking is going to be the next battleground of innovation and certainly transformation. This is your wheelhouse. You've seen this movie from old school networking to telcos, to now, we're in the middle of it. What's your take on networking because we're going to be at Cisco Live Barcelona next week but you have a good view into the overall landscape. Tell us, what are you seeing? Sure, no, I mean, I think you said it right. I mean, basically, if you look at what happened in compute in the last decade, plus or minus in terms of the advent of the public cloud and the cloud model and automation, even if you're using private cloud, you know, that is all going to, what happened last 10 years there is going to happen in networking in the next 10 years because, you know, there's no reason why these things aren't going to happen. The cloud has changed the model. People work, people expect to be as a service. More and more of the network traffic is moving from the enterprise into the cloud or to reach the cloud. And that's changing how, what the important isn't the network, it's changing how the network's going to operate. The cloud certainly has been game changing and I want to get your thoughts on this. You call it, and before we came on camera I want to make sure we get this out, the cloudification of networking. Correct. This is a term that you coined. What does that mean? Tell us, what is cloudification of networking? Sure, well, there are different elements to it, right? But, you know, I think if you look, I think the first thing that's important is to look at it from the user perspective, like what does it mean from a user, right? So if you look at their classic network, you know, 10 years ago, most of your traffic is in your campus and in your own data center. And, you know, you're using standard Cisco switches, you know, you've got, you know, some hierarchical system, you've got your data center, you're using 6Ks in there, so nexuses, whatever, you've got your N switches, some Wi-Fi, that's what the world looks like, that's where everything goes. And you may have some WAN with MPLS and maybe things like that, but you know, that's not a big use or some internet, you use internet, but you know, that's not your mission critical network. If you now fast forward to 2020, I can't even say that, it's like- It's been 10 years. 2020 is like, you know, God's, you know. All of a sudden, from a user perspective, where is your, where are you getting your packets from, right? Most of your stuff in a public cloud. So your traffic, whether you're in an enterprise or at home is typically going towards the cloud. The cloud itself has its own very complicated network of 100 gig up to even getting going to 400 gig at some point connections, massive, massive data centers, fully automated, right? In the campus, it's all Wi-Fi, but again, most of the traffic is back and forth to the cloud. So if you look at the network architecture, the network architecture that you have, the cloud now has completely changed the network architecture from being mostly internal to all being from you to the public cloud and the public cloud itself now has become such a super compute system that it's like a massive computer. And of course, massive computer has massive networking. Now, when you think about cloud, so that's from a user perspective. Now, let's talk about what this actually means. Okay, so first of all, when you think about the cloud, okay, what has enabled the cloud? I mean, the cloud, there's nothing special about the cloud. It's a bunch of CPUs and coordinate jobs, okay? But it can happen fast. Why can it happen fast? Because it's automated, right? It's automated, it's all software, right? It's CI, CD, you develop, you deploy, it's boom, boom, boom, right? There's nothing you're doing there that you couldn't do 20 years ago. It just would take a year instead of a day because nothing was automated, okay? Now you look at networking, right? That means that everything is gonna be automated and software-centric. So if you look at the way the network is today classically, right? It's not really automated. Yes, people have done some scripting and so forth, but it's not an automated, dynamic environment that is software-centric. So if you look at, for example, the data centers that people are building, if you look at what the big public live guys have deployed, right? They've deployed a very automated system. You look at and say, well, how many sys admins there are per how many nodes, right? And the ratios in these bubble clouds are things that you've never seen in enterprise. So you're seeing the data centers becoming these very cloud-centric networks and then you're seeing the whole traffic pattern chains where you're missing critical network now is from where you are to the public cloud over the WAN. And this is what's driving the evolution of SD WAN. So essentially WAN used to be old school, you had your local, local area network. You mentioned the data center and PCs, all connected hubs and printers. And then you go out over the wide area network to something else, another building, another data center, maybe a third party cloud when it was starting. Now you're saying it's all essentially going to the cloud. So WAN is the new land. And what happens also is that you quickly get on one of these big networks. So you think about, okay, what does the WAN look like? Historically, the WAN was a bunch of links between data centers. You had lease lines, you had like every enterprise had I don't know, 20, 30 data centers worldwide and they'd be linked in the WAN. Now what you do is you get on the first connection into an Amazon or Google or Microsoft network and then you're in their network. And so these guys have massive networks. Like you ask me, who's got the biggest, most impressive WAN? It's Google. They built it for themselves, right? If you put a package- You're saying Google's the best network of all the clouds. Yes, yes, and it's not necessarily long. They don't have the long term, but Google has always been crazy about performance, right? They're crazy about performance because that's about how fast can you get a search result and so on. And they had, that's money. Yeah, I mean- Speed is money. Yeah, speed is money to Google. And so Google have from day one looked at ways to optimize their network and their internal network and the Google cloud gets the benefit of that, right? They are crazy about performance. So they have, Microsoft is getting better. They're working a lot on it. And Amazon I think is now, Amazon is very good inside their network. So they've built a lot of, they've been very focused on how fast they're external, like their data center between the servers and so on. You know, they did Anna Peruna, they do all kinds of optimizations there, which have actually led the industry. But when you go into the WAN, that hasn't been a big focus for them to date. Where Google- And Google's got great security too. They've had a lot of good security. So from an inside a data center perspective, I would say that Amazon is top. Once you start talking about the WAN connection, Google is the best. All right, so I'm trying to pull this together. So the cloudification of networking means, if I hear you correctly, is what the LAN used to mean with the data center. You had a data center and you had LANs. That was a campus or whatever. Yeah, yeah. Is now the cloud. The cloud is now the new data center and that's the WAN is now the new LAN. Correct, exactly correct. The data center of the cloud, the LAN is the WAN. So, okay, I get that. By the way, I agree with that. I think there's going to be a massive disruption. And again, there's a long tail. So you can easily say to me, no, well, look, look at all these LANs. Sure, yeah. Again, Wi-Fi. Future is here. That's not evenly distributed, okay? So there's a long tail, nothing's disappearing. But today, How many nicks are in the PCs on the desk these days? Probably not a lot of PCs and no nicks. Only Wi-Fi. It's Wi-Fi. And also, look, the question you have to ask is, how many data centers does the average enterprise have? How many new data centers are being commissioned? You know, when I was at HP, you know, a decade ago or whatever, right? We had data centers up, there was you. That's where everything was, right? Nobody commissions data centers. Now the only people buying data centers are cloud guys. Yeah, and we'll get to equally next conversation. I want to get to the outpost with Amazon because I think that hybrid edge is an interesting trend that will come into the whole network edge thing. But let's stay on this WAN thing because if you're saying, and I'm agreeing with you that what the land was to the data center, you have WAN to the cloud. So if the cloud disrupts and creates an innovative enabling opportunity for startups and the clouds to create new value, it's going to be there. So with that being said, if you believe that, what happens to the old SD WAN markets? Because the old SD WAN was a riverbed, it was a box. You connected one thing to another. Yeah, so, so again, there's a first of all, it's on paper, it's a revolution. And in the market, it's an evolution. Right? And so, SD WAN itself has been moving in different directions. It started off with this, okay, we do choices. Now it's more about WAN. And you still have the telcos who are offering services with boxes in front of them. And that's not a bad business, it's growing. But over time, the question you want to ask is if the Amazon or the Google or whatever network becomes bigger and bigger and bigger, how many hops do I need before I just get on their cloud? And then I'm just under cloud and they have a massive internal network. I love the expression. It's not your grandfather's blank, it's always been an expression. So it's not your grandfather's SD WAN would mean, like if you look at all the incumbent players, you've got Cisco, even with VMware, they all have SD WAN plans. But SD WAN's different now. So what does SD WAN look like? Because you have startups coming out, you have security companies, I'm covering the news on security companies that are becoming SD WAN players. Because the firewall and your WAN access is becoming the same thing. So if you're a Palo Alto, if you're a firewall company, or if you're an SD WAN company security and if you're a firewall company, you need to have SD WAN, it's becoming the same thing. The firewall used to be the entry point into your network. And now it's SD WAN because that's like a distributed firewall. That's your perimeter. And so now the difference, what's happening in SD WAN is, in the early days of SD WAN, it was about choice. Like you basically said, okay, I can put my slow traffic or my less important traffic on standard internet and I can put my important traffic on my expensive MPLS links, right? Problem is MPLS is not very dynamic and people want a lot more capabilities. And so now it's much more intelligent where you have various players saying, no, I can give you a faster link, not over, like Teradian, the company I used to work for, did that, as well as other companies as well. And so now it's becoming a lot more sophisticated than that. It's like, okay, give me the traffic and I'll figure out how to get it across. And a lot of people are saying you don't even need MPLS anymore. But the thing that's important to understand about SD WAN is- But that was lockdown, traffic routes. That was essentially the application workloads that were earmarked to be high priority. Now you have dynamic- Now it's dynamic, but I think the important thing that the question that you need to ask yourself all the time is, what is the goal? What are you trying, why does SD WAN exist? That's my question. So SD WAN, so why does WAN exist? So WAN exists because you want to connect to resources that are further than WAN, right? SD WAN exists because the existing connections were either too expensive or too inflexible. So we figured out a software way to do that, which is better, right? It's sort of, but if all you're trying to do is get to the cloud resources and the cloud is expanding it's going to eventually push on SD WAN. So I got to ask you, why has there been a renaissance in SD WAN? What's the reason? I don't think there's been a renaissance. I think that- It's kind of seems to be growing, but it's not a renaissance. It's been growing. I don't think it ever stopped. I mean, there was software-defined networking which was sort of all the time. Why is SD WAN exploding in popularity? It's exploding because of the cloud because what happens is, if you're an enterprise, all your traffic now is going to the cloud and you need a dynamic WAN to support that and MPLS isn't going to cut it anymore for you, right? It's not dynamic. It's not flexible. Even if it's not expensive, it's too complicated and internationally suspended. So if you have like offices all over the world and all they want to do is access your cloud, your sales force, and so forth, it doesn't work very well. It's not dependable. It's not a, unless you tie in a bunch of MPLS lines all the time, it's not very reliable. And so you want something that can do that reliably. And SaaS drives that, the cloud. SaaS drives it, correct. The consumption model of SaaS is driving the need for a better WAN. The current solution for a better WAN is different flavors of SD WAN. I want to ask you about Cisco because one of the things I've really been focused on is the evolution of Cisco. They own the routes. They have BGP. Cloud has networking, but some needs at the enterprises might have, certainly at the edge and with the edge of the network and Cisco has that position. The future of networking and the cloudification of networking, if we take this to the next level is puts a bullseye on Cisco. It certainly shines a spotlight on their market position and potentially opportunities or losses that they may incur from it. So there's opportunities, the scenarios of where they can gain big time and the scenarios where they could get flat footed. What's your thoughts on that position? That's a really good point. Look, the good news, so yes, if you look at what happened to the various incumbents in compute, the move to cloud was interesting at best for them, right? And so obviously the cloudification of networking and the fact that more of it will be in the public cloud is going to present some challenges to the existing network players, the biggest one of course, and probably the most capable of them being Cisco of course. But the change will be a bit different because networking, as all the networking geeks always say, is a bit different, right? I think the first challenge that Cisco faces is we're moving to as a service, right? And they've started to understand that and you have like Maraki that provides everything as a service, so I think they're just the business model thing, they have to figure out, okay, well, how do we charge? Because people are used to cloud, they're saying, well, I want everything as a service. I think the bigger challenge that I think- That's working out for some. Look, Nutanix, that took a year. It took a lot of funds in our stock price for a year, but the pivot moved. Yeah, they have to move. And again, I mean, again, you have to decide when you make that switch, the innovator's dilemma, right? Especially someone like Cisco's got massive business, you can't just say, oh, this is cool, let's go do that. That's great. So it's a difficult, it's obviously- Right, so they can do as a service, what else do you see? They can do as a service. I think that the bigger issue for them is when you start doing as a service, then the question is what assets do you own in the network on the WAN, right? Today, they tried to do this, when the cloud came up, they had their multi-cloud, I forgot what they called it and that really, they're in play out because they weren't really a compute provider to do that, so that made sense. I think in the network, they have a lot more assets, but I think they're gonna have to own some assets. I think to provide a network as a service, they can go two ways, they can rely on the telcos, but I think Cisco's a bit more agile than the telcos and I don't know that that's gonna work, they're trying to do that. They have a new chip coming out. Yeah, they're trying to do that to rely on the telcos to do this as a service, but I think ultimately they're gonna have to figure out how to own some networking assets if they truly wanna be a player here. Can you say assets, what do you mean by that? More assets- Own networks. Own a network. Like if you're looking at the, if you look at Amazon or Google and so forth, they're owning, they're building out networks. If you wanna do a network as a service and you don't own some assets, right? You're relying on someone else. You're relying on someone else and that works, but like Apple, if you wanna control the experience, you need to own it. Now, they understand, the good news about Cisco is Cisco understands all of this. They understand networking, they're very close to customers, they see what's going on, but they have innovators' dilemmas they gotta deal with. So I would agree, you think they understand what's going on. Yes. And they just haven't made their moves yet. Or move- They haven't made moves that are obvious and also, dilemmas are difficult, right? Innovators' dilemma, there's a reason why it's called a dilemma. If it wasn't a dilemma, then you just go, okay, let's do this, right? So it's easy to sit here and pontificate while they're protecting whatever 50 billion dollar business. I think that they understand they gotta make choices and they gotta pick the timing of when they do- That's the key point. Sometimes a best decision is not to move, but also the time that they waste, you can't get that back to either. So timing's everything here. What's your, in your opinion, if you were at Cisco, what would you do? Get some assets? I think that historically, when Cisco had to wanna make big moves like this, the way they got around waiting is acquisitions, okay? So I think that they will have to acquire some, if they wanna make this move, they choose to make this move, it will involve acquisitions, maybe deeper partnerships with some people who currently own networks. There are various players in that area, in that space. But historically, at least that's been their playbook and that way, because they have a big checkbook, they can make up for the fact that they took their time. Yeah, but they also, John Chambers addressed transitions with a lot of M&A and it ended up biting them in the butt a little bit because you got to integrate it all. I personally think that Cisco's opportunity is to be the on-ramp for multi-cloud, do the, own the edge. And I think they need to find their TCP-IP moment. If you go back to the OSI stack, that changed the game, that created internet, we're in that created Cisco, the open standards. And I see the whole world almost going to proprietary. I remember those days, you had Decnet, SNA, proprietary Nostas, right? And what's interesting now is everyone's going back to these proprietary walled garden kind of models, you know, open source with a twist. So the question I'm looking at is that, is there going to be this interoperable layer, abstraction layer that could exist between creating cloudification of networking and more value? Look, I think the opportunity, right? But again, that opportunity requires a big investment and Cisco has the ability to make that investment is to be the on-ramp to the multi-cloud. But that means your network has to be better than the cloud guys because you're an honest broker. And you have to have a map, but you have to have a massive footprint. Whether it's through partnerships or whatever, you have to have a massive footprint. That footprint costs money. So like I said, I think if someone could do it, Cisco could do it, but you know, they have the service providers who are their partners. So, you know, they may not be that happy. Requires a big investment, requires a different mindset. I think they understand the challenge. I think again, they have innovators dilemma, but it's an opportunity and a threat for them. Well, Sar, it's great to see you. I know you're just chilling out now, advising, investing. You're clipping coupons, sitting on the beach. Beats is hard, but you're doing some fun things. Let's get you back in. There's some things I wanted to chat next time. 5G network automation. I think the analytics auto tuning AI is going to be a big thing. I think Wi-Fi application awareness. These are topics I would love to have you come in and do a drill down on. Appreciate it. Always great to be here. Great. Sar Goli, friend of the cube. Now analysts here in our studio for future of networking, we're going to continue the series, more networking, cloudification of networking, a big trend, the cube is on it. It's going to impact the edge. It's going to impact enterprises. It's going to impact how we do business. And more importantly, how software is being built. This is theCUBE here in Palo Alto. I'm John Furrier. Thanks for watching.