 At Oracle OpenWorld 2012, this is SiliconANGLE.com's theCUBE. theCUBE is our flagship program. We go out to the events and extract the signal from the noise. We go to all the events we possibly can and where the action is, we are there. SiliconANGLE and Wikibon are on the ground providing great coverage, blanket coverage and breaking analysis. I'm John Furrier, the founder of SiliconANGLE.com and I'm joined with my co-host. I'm Dave Vellante at wikibon.org. It's our pleasure to introduce you to Lee Doyle, first-time CUBE guest, a friend of mine, worked with Lee many, many years. Lee, welcome to theCUBE. Thanks, Dave. Great to see you. Yes, so Lee and I, we work together. God, I think we started in the same week at IDC back in the early 1980s and... Is it that long ago? We don't look that old. So Lee, you know a lot of people in this business. You've seen some major changes over the years and the hot thing now is software-defined networking. So help us squint through that. Is it, you know, what's real? What's hype? What does it all mean? That's a great question. It is the hottest thing going on in networking these days, for sure, and it is real, and it's going to be a change in how networks are architected and how agile they are and how they're managed and how you architect for the cloud. So what does it all mean, software-defined networking? How do you define that? So it's really the ability to abstract the intelligence and separate some of the hardware from the software or the control from the data plane is the classic definition. But a more real definition to me is having open APIs or having an open ecosystem around the network as opposed to simply being this closed black box where you optimize software for ASIC performance. So networking is an interesting industry from the standpoint of Cisco has dominated for such a long time. I don't know what its market share is. It still is. Well over half, right? It's two thirds, you know? So is this trend, the software-defined networking trend, is it disruptive to Cisco? Is it going to allow others to gain share or is it more evolutionary and Cisco will just continue to dominate in your view? I can go a little bit in both but it is potentially disruptive because it does offer a new way of doing things and those who are more nimble and who can move faster and the incumbents always tend to move a little slower in these markets. So what's your take on Oracle Open World? I mean, is this a show you've gone to frequently and Oracle's not really a networking company to say but so what brings you here and what's your angle on Oracle? Well I've traditionally worked with Oracle on the telecom business and they sell a lot of servers and other things to the telecom industry but also networking's a big part of the cloud story these days and so I expect Oracle to embrace software-defined networking at least in part. Well that's their MO. They tend to go after the trends that kind of percolate and kind of cross the chasm and then put their blanket around it and bring it in, wrap it into the fold and vertically integrate that Larry to get his data center future. It's no secret in the industry for folks who have been in the industry long as us and you've been covering the sector. And Larry wants his own data center. He's always dreamed of being HP. He's had HP envy for years back in the day when HP actually was envious in terms of like the Bill and Dave days but now with HP struggling he's obviously got Mark Hurd on the team. He could get there. I mean he's got the package. So how do you as an analyst look at that and say okay we got to get some networking in there but hey Larry you got a multi-vendor situation going on. It's been going on for the history of the tech business. This is not a pure play consumer clean sheet of paper. You got legacy, you got multi-vendor, you got interoperability. That always comes around in every cycle. It always comes back, interoperability. What's your take on that and how does Oracle need to deal with that? Yeah it's an interesting question. I mean there's clearly the ability to buy a company. There's lots of startups and VMware has taken its plunge with 1.2 billion with the CERA. There's the partnership route and there's also open source or an open community around open flow so there's a way to play it all of those different ways for Oracle depends on what they want to do. What about the confusion, the market? I'll see the Nacera deal pops everyone up and say hey you know something going on here. This is important, yeah. Wow I mean big step up from valuation when they just did their last venture round now that's a billion dollars and so everyone's kind of high-fiving each other in Silicon Valley but it puts everyone in a notice. So at VMworld when we had theCUBE there we noticed an air of a cold war. We called it the Cuban Missile Crisis of Cloud where Cisco steps back. They got this weird fresh release with VMware. Got Juniper kind of going wait a minute we did Junos and you got open flow startups. So you got this like balancing kind of a re-tweaking of everyone kind of a reset. Right. Explain that to folks out there. One is that real or is that just more a facade of the dynamic? And what are the big postures of the big vendors? What are they doing at post Nacera acquisition? What are they thinking? Right, well I see three categories of players. So there's the startups. There's Nacera, now VMware. There's Big Switch, there's M-Brain. There's a lot of other guys announced and unannounced in the valley. So there's a ton of money there. There's the networking guys and they all have a transitional story, buy our stuff and we'll take you to the SDN vision that includes Cisco, that includes Juniper, that includes Brocade. And then you have the IT players which include VMware but also HP, IBM, Dell and others who need to have a full Oracle perhaps who need to have a full stack. So if you software define data center or however you'd like to define the next generation cloud build networking has to be part of it. Yeah, we gave you a little peek of our software led infrastructure position we're coming out with because that's undefined. So software defined data center really means nothing right now. Which is good, that's a marketing dream. So marketing guys out there scramble and put stakes in the ground but the geeks call it network virtualization. The market calls it software defined networking. So connect the dots. Where is those next journey? What happens between network virtualization, software defined networks and then the destination of software defined data center. Is it the same thing, is there tweaks to the story? Well SDN is really about helping the network keep up with the migration and the tremendous innovation that's happened in servers for sure and then storage to maybe a slightly slower extent. So you need to provision virtual machines, you know to move them, you need to do so securely and you need to do so in a very high performance way. So that's what software defined networking is primarily about. It's about public cloud builds, then private clouds and then we can talk about campus and branch and WAN and all sorts of other things when we get there. So I wonder if you could put your practitioner hat on for a second and talk to the users in our audience. What does all this change mean to the user organizations out there? What should they be thinking? Right, well you know if you're looking at the cloud holistically, let's say it's private cloud, moving to a more agile data center, right? You've got a number of different parts in that. In your networking part, you need to look at what are the APIs, what are the ecosystems, what are the tools and some of the software that you can use to leverage software defined networking. Because networking without tools without something to leverage it, it's just more networks. You need to look at management and provisioning. Ideally you want to reduce OpEx. The manual configuring of the network is just very time consuming and very costly in people terms. Security's a big issue. Also from a legacy standpoint, everything is you have to migrate from. Unless you're starting over, unless you're blown up or going completely green field, you start with switches and routers and go from there. And the plumbing's not going away. So you want to have a high performing, resilient network architecture as well. Lee, what's your take on the whole service provider trend? It's interesting, you go to most of these events and certainly at VMworld, you heard a lot of focus on service providers. We were at EMCworld, we were at SAP Sapphire. There's a real discourse from the host vendor to really try to court these big service providers. You haven't heard a lot here from Oracle. Now maybe they're selling technology in there, but certainly not from a partner standpoint. They're not hearing, oh, we're going to partner up with the service providers and deliver clouds and they're going to buy our technology. What's your, two questions here. What's your take on the whole service provider trend? Is it disruptive? Is it new form of distribution channel? Is it just the evolution of the services business? And what's your take on Oracle's angle with service providers? Well, I think it depends on how you define service provider, right? Being a networking and telecom guy, when I think of service provider, I think of AT&T or Verizon. But you could also think of Google and Amazon and Rackspace and a whole lot of other folks. So clearly there's a whole lot of servers going and storage going into these organizations. So they are a new channel. So they're a buyer. Yeah, a huge buyer, right? And the traditional telecom companies really need to change their business model and are changing their business model to be full scale cloud providers, especially the tier one folks. Oh, we were talking off camera about Dell a little bit. They've really stepped it up in a number of enterprises, certainly storage. They've always sort of been in the server space with the x86 service, but also networking. They've made some acquisitions there. What's your angle on Dell's networking strategy? Well, networking and security. So they bought a number of companies in that space and they're trying to put that together with their server and storage to really go to traditionally mid tier enterprise, right? Is the Dell story. So they're trying to package that all together and leverage that into a consistent integrated offer. Do you consider them a major player in the space that you follow, Dell? I think that their player in the networking area, are they going to go after the big public cloud? I don't know. Lee, talk about HP because we're bullish on HP. We've been critical of HP. I have been in and taken some heat for that with HP, but not the audience, because I love HP. I've been working there nine years myself. If folks know that, I've been very transparent about that. But they're networking group solid. I mean, they've had a great legacy in networking, going back to the ether switch days back in the 90s. And then they bought three coms most recently. They just solid on networking. They've been shipping an open flow product. Bethany Mayer was on theCUBE talking about that. And they're not pumping it up. So they got some work on marketing. Like Bannex and others got to do some marketing on that. But for the most part, they're solid. They're part of that ESSN group now called Enterprise. So they've been executing converged infrastructure in a cool way before anyone else really is trying to go over there. What's their current situation right now? Are they just not getting the word out? Are they being smothered by the HP kind of reboot? And what's their prospect? Give us your take on HP at the moment. Sure, I'd love to. They're the number two networking vendor behind Cisco. So they've been growing in networking. So they're a solid player, have quite a broad product line and people respect their products. I think they do suffer a little bit of the HP hangover from the corporate side, but they've been busy getting the message out and putting together a data center story that leverages the network. They have a good story? They do, yeah. How about the servers? Do you follow the servers at all with that group? Not, you know, a more networking guy. Well, what about the converged infrastructure play? I mean, that obviously affects networking. It's kind of an evolutionary trend, but people are using that as a wedge, as a way to compete in the marketplace. Clearly Cisco getting into the server business looked like it was going to be disruptive, but it seems like it's more just an evolution now. Have you tracked that market and what's your take on it? Right, everyone's trying to put together server storage networking with the management wrapper and make it a lot easier. I think HP has done some really interesting things there, as has IBM, as has Cisco, so there's a lot going on. I don't think it's all been played out yet. Are IT organizations ready for that convergence? Are they organizationally aligned for that? They're thinking about it, but are they there yet? Probably not. We had Jay Sri, the CEO of Arista on theCUBE. Many times, she's a dynamo. We love her, a big fan of her work, and also she's great on theCUBE. She gives great content on theCUBE, but Arista, they got funded. They got in the trenches when there wasn't a lot of funding, and it wasn't cool to be a networking player, and we've talked to Doug Gore-Lay many times. He's a friend, he's been at Ex-Sysco. It's hard to build a systems company, and I talked to the founder of, Pradeep, founder of Juniper. Doing the networking startup wasn't like it used to be. It's really hard, it cost a lot of money. So Arista did it when it wasn't sexy, and now SDN hits the table, and they're kind of clawing, hey, we're here! So, and they have a good product. I see their bosses in all the top data centers of the web-scale companies. What's your take on Arista? I'm relative to, one, their current situation, and then now the market's spinning with SDN. Well, I wanted to comment on your startup question. So software-defined networking, you need tens of millions of dollars, maybe even a couple million dollars to startup, as opposed to, you know, Arista and Juniper, they need hundreds of millions to build the A6, right? So order of magnitude, so there's going to be a lot more software. It's not a super angel deal. Right. This is not a micro VC, 500K, put up a website. Well, Arista has built some very high-performance networking, they've got a niche, especially at the high-end, and they're now starting to tell their software-defined networking story. And Jay Sherry's done a nice job of putting some of the comments I've seen there. So you talked about building A6 and some of the custom silicon. Yet, we are living in this world, everybody says it's going to commodity, software-defined implies commoditization. We had Jay Sherry on, one of the first things out of her mouth was software-defined, and she's very savvy, taking advantage of those trends. But what's your take on that? Are we moving toward a more, you know, a less custom A6 world, or is there a more of a demand for that type of capability because of the networking bottlenecks that we're seeing? I don't think software-defined networking doesn't presuppose commodity. It's an option, right? There's always going to be plumbing, and plumbing, you're always going to need high-performance plumbing, and high-performance plumbing is going to require A6. Now, that being said, where's the balance of intelligence? Where's the commoditization? Is there going to be more cuts? Is there going to be more off the shelf, you know, x86 in networking over time? Absolutely. How fast and where? That's a great question, and one that I've been looking at really closely. So we were talking about NYSERA and VMware before, and Cisco, and of course Cisco and EMC, John Chambers and Joe Tucci have this very tight relationship. VMware goes out and buys NYSERA, everybody says, oh, that's the end of the relationship, and of course they, in public, have, you know, great relationship. They're doing a lot of business together, but if you fast forward five, seven years, you can see where a company like VMware will gain more networking function into the stack, try to commoditize the network where possible, because they want to keep it as low cost as possible, but I guess your argument is there's still room for guys like Arista and Juniper for the highest performance applications and use cases to tuck in there, but what's your take on that whole trend? Right. Well, you still need the network to move the bits, right? It's what's controlling that operation and where's the value chain, where's the profit pools go, right? And is VMware and Cisco on a five-year collision? Absolutely. So let's talk about Oracle. So Larry Ellison said in the keynote, small little comment, of course we picked up on it because our name's Silicon Angle. He said, Silicon, we want to go with Silicon. So we all know the Apple move, right? And I know this because a bunch of my friends were engineers for a bunch of stars that Apple ended up buying for the Silicon. Right. Apple made these moves early on, buying guys, building chips, proprietary chips, purpose-built for specific hardware functions. They have Sun, Larry's got to have his own chips. That's in the future. So where would you see that fitting into them, obviously, to this day? Any perspective there? We're trying to tease that out because no one really picked up on that comment. I mean, it's hard to even get a, it's hard to tease Sun out of Oracle these days, right? But it's still a very important part. It's now fully in there? Well, also, you don't hear a lot of talk about Spark. Is that because Oracle was a software company that bought them for a yard sale? Or Larry's got his own little hardware group? I'll let you draw your own conclusions there. I think I just said it. But Sun is a toy for Larry. He's building it. He's not blowing off Sun. He's fully putting it in there. Right. I mean, that's a core part of his strategy. Well, I think a really interesting aspect of this, and one of the things I'm here to learn about is how Oracle will take its engineer-assistant approach and really build that into the networking and telecom industry. Because that's always an industry that has had very highly specialized hardware and highly specialized software in an optimized black box type function. You don't just split it out. Right, and you're seeing some adoption of these converged infrastructure solutions within telecom, and in a way, Exadata was one of the first converged infrastructure solutions, although it's not a general purpose converged infrastructure solution. So, Oracle, if I understand it, really doesn't have a clean product play there yet. I'm sure they have a lot of different products that, you know, Sun Legacy products that they sell at its telecom, because Sun had a huge telecom business, but what do you think the right move for Oracle is? Should they go, for instance, to go compete against the VBlock, for example, or some of these other reference architectures out there like you're seeing with FlexPod at NetApp in the partnership with Cisco and VMware, or will Oracle, in your view, just take a typical Oracle different approach? Well, let me take a slightly different angle on that question, which is in a little notice acquisition, they bought a company called Go-Head. So Go-Head does HA middleware, which is SAF compliance, which is a technical way of talking about telecom-specific middleware. And so they're going to bundle that in with their engineered systems and really try to offer a platform for both service providers and network equipment companies to build their specific applications on top of that. And then we were talking off camera as well, Oracle bought Zygo and they made it a, they marketed it as a software-defined networking play. It's really an IO virtualization play, you know, a form of consolidation, virtualization. It's like Oracle's picking off these little pieces and then they end up putting them together. And next thing you know, they announced the Oracle Public Cloud or the Oracle Private Cloud, and it's all red stack in there. It's a really different strategy than you see. It's very IBM-like of the 1980s, early 1990s, isn't it? Right, and they can clearly offer the full public cloud or private cloud stack to the telecom industry, and they clearly intend to. Right, how about mobile? We haven't talked anything about mobile, but have you been tracking mobile and its implications to the networking business and what's your take on that? Sure, you know, BYOD, the wireless LAN, and once you have a network that is wireless-led, as opposed to wired, now you have a different architecture and you have potential disruption to go to a more software-defined arena and to look at what your incumbent is, because the plumbing's still in the background, but you're not as concerned about the actual hardwire plug. BYOD, that's a real challenge for a lot of IT. It's a management necessity, but how you implement it in a secure manner and a high-performance networking solution is not intuitively obvious all the time. Well, yeah, you've got a lot of inertia with the legacy desktops and devices and processes, and all of a sudden you have to support these new devices, it's not trivial. And oh, by the way, you need to build an app store. I mean, we really have anywhere, anytime, any device, is upon us. We talked about it for a long time, now it's here. All right, Lee Doyle, well, listen, thanks very much for coming by theCUBE, it was always good to see you, good insights, and I'd love to have you back at some point and enjoy the rest of the show, and thanks for watching everybody. We'll be right back with our next guest right after this.