 at SiliconANGLE.com. This is theCUBE, our flagship program. We go out to the events, extract the ceiling from the noise, and we are digging into Dell world because all the hottest trends are happening. Can the infrastructure level, at the application level, at the mobility level, and everything in between for the enterprise, the consumerization of IT. It's exciting to dig in, and I'm John Furrier, again, John Furrier, the founder of SiliconANGLE. I'm joined by my co-host. Hi everybody, I'm Dave Vellante of Wikibon.org, and we're here with Darius Samarian, who's the Vice President and General Manager of Dell's Enterprise Systems and Solutions Group, new group at Dell, really designed to go after the converged infrastructure, market opportunity, which we were the first to define that, and in size that, it's huge. It's about a $400 billion TAM, so Dell is responding to that organizationally. Darius, welcome. Thank you, good to see you again. Good to see you again. I'm glad you're here. Yeah, we saw you this June at the Dell Storage Forum. In Boston, yeah. At the time, you were the networking guy. I was. Now you've been promoted, congratulations. Marius Haas is in. We know Marius from his days at HP, and even before that. So you guys are really going after this converged infrastructure play. Talk about the organizational change and what drove that. Yeah, very much so. Pleasure to be here all the time, John and Dave as well. If you think about what we've done over the last three to five years, the whole idea of arguably spending $12 billion worth of acquisitions is allow us to accumulate the necessary technology assets and people capabilities to finally be able to culminate into what we can call system and solutions, orientation for Dell next generation. The reason why I was a logical perhaps candidate to do this is that I came from the networking editors where by definition, converging for structures become very much the next conversation with our customers. And so the formation of this new business unit called enterprise system and solutions should be seen as a, call it a P&L that sits on top of our server storage and networking because the combination of the three plus of course system management and plus the collaboration with our software group can give a set of capabilities that before was not available to us. And right now by formalizing this, we can actually talk to our customers even more straightforward and being able to have a conversation around what really matters which is the application, the private cloud and the infrastructure being built much more integrated, much more converged than ever before. So Dari, I want to ask you, this is really an important topic to us because it was one of our coverage areas, big data and now software-led infrastructure. So Wikibon put out the first groundbreaking research around what we're calling software-led infrastructure which is a way for us as we've been doing the research to modernize the definition of converged infrastructure. So one thing that we've observed the customers that you guys are serving as well as your competition is, converged infrastructure was defined many years ago. So a lot's changed in the past just three years, okay? So as you guys are transforming the messages, obviously that you're going after this modern infrastructure as Michael Dell kind of responded to my question yesterday around modern, I wanted to drill on that and that was a word he used and we use that as well. So we have this notion of software-led infrastructure. So share with us within Dell because it really is a nice architecture. Obviously server stores networking, they all have to work together and software plays on top of it. But software-led infrastructure, talk about this modern definition from your standpoint. Obviously we have a report, you go to wikibon.org, you find that, but this modernization trend is happening. IT and CIOs are all looking at their infrastructure. They are looking at cloud. They are looking at software to be differentiated. And obviously this margin expansion in there too for the people who are providing those durable goods. Sure, sure. No, very much so. In a way the definition of a modern data center starts first with recognizing that the real value that customers want to get out of a data center is the running of the applications. And you can run applications either on a very confined basis and fine tune your hardware and your software system management for those. Or you can actually go and create a private cloud where you have a method to be able to take advantage of multiple applications sitting on top of a private cloud. I will argue that the advent of a modern data center was really triggered by virtualizing the whole data center itself, the whole server stuff. And what is being happening is right now that the virtualization applied to servers when it catches on as we are doing collectively in the industry to the storage and the network side creates an interesting way to disintermediate in a way the underlying hardware and the application sits on top. And our job as a solution provider is to be able to provide the best possible optimized infrastructure that allows to have a virtualization way to sit on top but all at the benefit of the applications that the customer wants to put on display. And a key topic is around how do you work with a customer to actually recognize that they are typically in brown field environments and they don't start from scratch. And we have developed a whole set of methodologies with our professional services group to be able to actually move the customers from the legacy applications into a private cloud based data center and our technology efforts helps in doing that in a much more economical fast. So what you're saying here, the old way was infrastructure application had middleware, I was kind of the old model, monolithic in servers. That's cool, but with this new way virtualization really is enabling this new kind of software layer, if you will. Okay, so that being said, talk about what that means. Okay, so we see the hottest trend out there is software defined networking. You know, VMware, bot, and Acira, there's all kinds of conversation around hypervisor, agnostic, so the virtualization in particular, and what does all this mean? What does software defined networking mean? What does software defined data center? And what is that trend? Is it just virtualization or is it other software playing with the sandbox? Fantastic question, in fact, the word virtualization and the whole phenomena virtualization, I will argue is only the trigger that makes all of this possible. So let's dissect a bit more about the notion of what a software defined data center, for the matter, a software defined networking is all about. I will argue that the value to a customer is about how dynamically the network and infrastructure can be customized to the needs of an application. So if you think about the essence of software defined networking, the way we are going about it is to say that the application has certain networking needs. It could be bandwidth requirements, it could be quality of service, it could be security requirements, it could be opening VPN tunnels, whatever it is, our ability is to actually profile the application, use a controller in between to be able to dispatch the instructions to the underlying network so that the application can have the most optimal network for the operation the application needs. If you do this across both the network, the storage, and the server, now you have an opportunity to program your data center in a real-time, as real-time basis as you can, given the needs of the application that sits on top. That makes sense. So you made the Force 10 acquisition some other smaller tuck-ins. Do you have all the pieces now to compete in this marketplace? Well, you never hold the pieces because by definition, technologies continue to move. But I will argue that we have a lot of the pieces that were required to go to the next wave of system and solutions. The whole idea for us to have about Force 10 as an example, it was an absolutely stepping stone to have enough critical mass of data center networking. And our ability to have built a very powerful blade-based switch that now connects to a very powerful top of rack, and it goes to the distributed core, you'll argue it is the most economical and the most elegant way to build a data center. Without that piece, I will still be immersed in a way of other networking vendors to be part of the overall solution. But even gale technologies that were actually acquired just on November 16, and we closed, it was an absolutely necessary step to be able to provide a workload orchestration on top of the iPervisor and to be able to do conversion infrastructure management. So if you think about this right now, the 12G servers that we've been all along, the progress we made with the storage site and all the progress we made on the Force 10 for the data center and the gale acquisition, now you have the building blocks, the elements for a conversion infrastructure proposition that we're going to put a lot of imposition. Now, one of the, John mentioned our report on software-led infrastructure. One of the findings we had was, you know, the commodity piece, you know, services on top of commodity hardware. That's really nothing new. I mean, people have been thinking about that. Certainly Google's been doing it for a decade, but the key is the metadata management. You've got this metadata locked in all these silos. What's Dell's software strategy with regard to allowing what you said, that dynamic reaction to customer needs? No, very much so. In fact, you saw yesterday in the Q&A session that even John Swains and Michael Dell had where John spoke very specifically about the ability to correlate events and correlate data across multiple areas of your software stack. Imagine if you can have a slowing down of a CRM application being correlated to the fact that there is a particular switch not being functioning anymore. Now you have a complete visibility of your data center and you'll be able to actually move around and being able to even proactively look for the optimization of the data center and that is only possible when you have quest on one side, when you have identity management capabilities on the other side and you have security and et cetera coming together. So the correlation of the trends and the correlation of the data in the data center is allows for an optimal, ongoing applications running on the other side. Machine is managing that. Machine managing that as opposed to the human being managing it by all means. So Dario, let's get back to the application. I want to kind of drill on that because obviously the application explosion is a real trend and everyone's seeing it at many levels. One is the agile ability of DevOps. I mean you guys are doing a lot of work in DevOps which means essentially there's more applications, more diverse applications being developed faster and deployed faster. All that's great stuff and we cover that. But what that means is a couple of things. So I want you to ask a question about three things. So there's been some innovations and some areas of improvement that people are working on. One, flash. Flash has enabled architectures to be different, caching layers both on the sand side and on the storage side and also on the server side. So I want you to talk about what flash means to the enterprise. Two, the networking side. Obviously networking has been a bottleneck around IO latency and dealing with provisioning of apps especially within a virtualized environment. And three, with the applications being so promiscuous in this consumerization experience, is the security holes everywhere. So three things. Flash, the networking bottlenecks being overcome from whether it's automation, configuration, and all that and then security. How do they all play in? So can you talk about all three? Yeah, let's try. These are big topics as we all know, right? I'll get 30 minutes of that panel. We could spend 30 minutes on this. We could spend probably hours on this topic. Let's think of flash as a way to arguably revolutionize how data becomes available to the application in a real-time basis. Meaning that if you think about, there is really the optimal way to distribute on a tier basis the data. And you can say, when do I put it off-premise? When do I have it in-premise? When do I have it close to a Nardisk? How do I have it close to a flash, close to a memory? Well, that is the storage subsystem, but all of this transformation and having a flash-based system is primarily to make available to the application, the data, as fast and as locally as possible and making sure, therefore, the infrastructure continues to pay attention to what the application is. So in other words, I would put the flash revolution in a way that would be possible. That's the acceleration of data, which plays well with the tax. It's the acceleration of the data and I call it the locality of the data, whereby the application would have access to the data at the right time, at the right place, and making sure that there is minimal latency and minimal, so to speak, delays for the data access. So would you say that's table stakes in the future? I would certainly think it's table stakes in the future for the simple reason that the application is going to become more and more demanding, more and more data is going to be there, and therefore, connecting it off between what the application is and what the data is, is going to become much more stable stakes in architecture as well. Yeah, we agree. I'm sure you do. But you know, it's good that you guys are on that because that's where, that's a bottleneck, but it's acceleration at the same time. So- It's very much so. But that's part of your core strategy. Flash in there. Very much part of the core strategy to have a whole set of advancements on the storage subsystems we put together for the active infrastructure. The second topic that we discussed for a year was network and this notion of bottlenecks. I think that with advent of 10 gig at the server level, and then 40 gig becoming more and more for the fabric itself, I think we've seen very good advancements whereby the optimization of the east-west traffic, whereby virtual machine stuff, virtual machine way more than going in and out to the data center has brought up this new architecture that you can call distributed core. You can call- This is software, this is the software-defined- This is both the architecture to which you have a spine and leaf on why you actually scale the edge of the network that is a physical side of it. And of course we work with the usual suspect on the silicon, you know, merchant side, you know, the Broadcom, the Intel and so forth. But equally important is the software that powers this scaling of the backbone of the data center network. We're working on that. Provisioning automation. All of those elements, i.e. the wiring up front for 10 gig going to 40 and being able to scale, the ability to actually provision it quickly so a switch can be provisioned as easily as a server, the ability to actually have automation kicking in so you can actually have open flow standards and as the end, that's the whole changes and evolution that we have in becoming a virtual network. The third area was around security. Well clearly, you know, none of this, you will argue, can be to the satisfaction of a CFO, a CIO and a CEO, you know, without having enough elements of security to be able to protect both from a liability point of view, but even more importantly, the applications having to have, you know, both the infrastructure security side as well as the end user being protected for the episode on the internet. And thanks to acquisitions that we made like in SonicWall, in SecureWorks, not only we have the architectural element with SonicWall, but we also have the monitoring element with SecureWorks where we can actually help our customers put together data centers or end offices, branches and so forth and protect both from the outside threats coming from the internet, but also securing a virtualized data center for the modern data center. So that's the core of the edge. That is exactly- That's what Marius was talking about yesterday. It is the core and the edge and he is the end user and the applications. All of these are the elements of a security architecture that we can bring to our customers now. Darius, I want to ask you about the organization. We started off talking about this new organization and now my contention has always been that you got to get the organization right, not only on the vendor side, but also the users. And I saw this was VCE. They didn't have the organization right. IBM itself didn't have the organization right. They just put Andy Monchon in charge. I would say Dell realizing you've got to have a leader across these server storage and networking areas to integrate these. Now, so that's good. Nice, good move, smart move. How does it work? So you're a virtual P&L. You got to be careful about double counting and I know you guys don't do that. So can you talk about that a little bit? This is my first question and I want to talk about the implications for customer organizations. It's a great question because you're absolutely right. If you don't get the organization right, all of these might be in generation and kind of waste of time. Think of this as the following way. There are two ways to make new products or new capabilities. One is a short-term ability to bring together what we got. The fact that we have a fantastic proposition on a data center, the fact that we have a fantastic regime on storage and servers, you want to take advantage of that and you can call that an integration job at the beginning. My number one job to start with is to actually bring those together and work with the other general managers and to put those in a way that are pre-integrated, pre-validated and easy to deploy and easy to consume from the customer. Job number one. Organization I think is structured that way and have an opportunity to sit at the table as in a way you can call it the chairman of a architected set of conversation with our peers. But then the opportunity is as we evolve is to actually have focused investments for new form factors and new architectures. They may not come logically by just summing up these three areas but that's to be done from scratch. My charter with this newly formed enterprise system and solution is to do both. Integrate best that we have today because the customer already likes the proposition but also go after organically and inorganically of the new integrated systems that actually may have to be developed and funded in a separate base. Okay, so you measure that but now the revenue flows into the individual groups. Revenue flows actually in two ways. First, I want to make sure that we can motivate our sales individuals and we can work with our customers and measure the units of pre-integrated systems. So very much the PNL that is a integrated system and solutions will measure both the pre-integrated system as a unit of revenue as well as the reference architectures. And then behind the scenes we do the accounting translations so I can have a correlation between the revenue that goes into network and the APEX spend on networking. But we don't want to have that accounting on the back end to be on the way or now we put together a solution for the customer. So this is very much not what I call a shadow PNL but a real PNL because it creates the right conversations and the right mechanism of execution even downstream with the field and with the channel. But real PNL in the sense that you can allocate investments according to the opportunity. That's exactly. Okay, so now what about the customers? Now the customers aren't organized. Generally, most customers aren't organized to exploit converged infrastructure. They still got storage, server and networking people. Actually some of your customers might be. Because the smaller customers might be in the better position. What's your handle on that? Very much so. I think what we notice on the customer base that first of all there is a big variety of customers. All the way to perhaps the most sophisticated, global 500, all the way to the mid-market and even small business. So having recognized the level of profiles of a customer what we typically notice is this. The most successful way for a customer to modernize the data center to bring converged infrastructure is when they come with the predisposition of changing their organization or boundaries entirely. When you have somebody that leads within the customer side a virtualization project or a data center project. Typically that individual, that executive that most likely reports directly to the CIO and reports very much higher up will be able to actually transform the company within because then you don't have so to speak the political infights between the networking admin and the storage or the server admin is somewhat odd. So we notice that the maximum success for us to go in and speak about systems, solutions and converged infrastructure is when we find the frame of mind on the customer that says I'm ready to also redefine the vision of labor on how my IT admins actually come together. Now you've talked about both pre-integrated systems and reference architecture. So you offer the spectrum because there's always pressure in the channel to unbundle, I want to use somebody else's servers or somebody else's storage. You allow that. We allow that for a very simple reason. You know, there are two kinds of realities here that we never could never neglect. Number one, every customer is somewhat different. Yes, our objective on a pre-integrated system is to tap into the 80% commonality and say, look, we've seen many like you and although you have a bit of a difference why we take advantage of our experience and our knowledge that we accumulate by talking to a lot of customers that you can call it is a pre-integrated system. But we also know that a customer may say, look, I really spend enough money with somebody else's to do our storage subsystem. Or I want to make sure that I preserve some switching from an alternative networking vendor. We don't want to neglect the opportunity that we can evolve with a customer. So reference architecture allows for the recognition of other components. So this does not become a forklift for the customer. It becomes an evolution. And therefore you can go all the way to a pre-integrated system if you like or all the way to evolution of reference architectures to have the customer journey coming along with us. Now, why the new branding from VSTART to active systems? VSTART's still around. But what was behind that? What was behind it is, in a way, the recognition that this is going to become and it still is a huge emphasis for them moving forward. We pick the word active as a label, in a way, to bring together a bunch of properties around the flexibility of it. The ability to actually make your hardware and infrastructure alive for the application. What we didn't want to do is to be bounded for some form of legacy. And therefore the word active infrastructure, active solutions, active system, active system manager is a whole concerted effort to bring together all the pieces of the puzzle. And therefore moving forward, active infrastructure is the umbrella marketing. But then downstream, we'll activate this as we put together the solutions for the sales, for the channels. We'll see us doing marketing around this. And therefore it is an opportunity, frankly, to bring together a number of pieces that together before we were not being as aligned as right now. And you see us doing a lot of, you know, VSTART stays. VSTART stays, but it will be evolved so that it becomes an active system as well. Into the umbrella. And the primary element is that when you have an active system manager, now the underlying, so to speak, you know, existing products that we have become part of the family. And the customer can be future proof from an investment point of view that says, I bought VSTART, why don't you bring me along on the active infrastructure, whatever position you do that by allowing them to adopt an active system manager and now they become all active systems. Dario, I want to ask you my final question. I know we're getting up on time here, but obviously a very important area for customers right now is this reconstruction of the data center. When I say reconstruction, I mean, it's a build out. And on the consumer side, you say you're in the Dell ventures and the consumer side's kind of, you know, thinning out the bubble burst, so to speak, in a way, good bubble, but there's a real boom in the enterprise right now. Venture capitals investing in flash startups, SDN, software to find anything is getting a lot of traction. It's obvious people are seeing this transformation is real. Even though we're looking at a financial cliff in 19 days with the here in the US and abroad, there's real work being done. So I want to ask you, how are you marketing this group and what's your main horse that you're riding into the customers? Is it software led? Is it software defined? Is it just converged infrastructure? It's kind of an old definition. Is that working? Is it big data? How are you going to your customers and what are you positioning? What's your core positioning? Is it software to find data center? No, the leading conversation with our customers starts at a higher level, to be honest. It starts at understanding what do you want to do? If you want to create, let's say, a collaborative environment and the customer has hundreds or even dozens of users. How do you put together, let's say, a Microsoft collaboration suite as an example? And so the conversation starts either at the application workload and what do you want to do with it all the way from moving from a legacy to modernization and bringing in the underlying converging infrastructure solutions or even a private cloud as a way to solve that kind of problem. All typically starts from a different point of view and says, I know I'm going to run a lot of applications and want to modernize my data centers because I want to spend less consumption, higher density, simpler operations. And therefore the second way to start a conversation is through a private cloud conversation and bringing converging infrastructure and pre-graded system under that, call it slipstream of conversations. So typically the customer doesn't ask us, tell me more about software to find networking or tell me more about software to find data center. We talk about it as a way to demonstrate that there are properties around it typically is the application of the private cloud that drives the conversation. I'll ask a different question. Obviously converge infrastructure is kind of an old definition. It's kind of played in the mind of the customer. Yeah, they're all been investing and now you have this big push for top line growth and bottom line reductions of being more efficient. So efficiency, which is doing more with less and making it all work, but also with apps are driving revenue. So the revenue focus. So is it like we're dels differentiating itself by saying we're betting on cloud. We're also betting on, is it just converge infrastructure? What's the core positioning that, when you go to the customer, who dels better because of... The core positioning with the customer we have right now is this. The customer is in different stages, depending on geography, depending on size, depending on the legacies applications of the transformation of the team for structure. And they all know that because they read the papers, they talk to their neighbors, they talk to their mentors or whatever, that there is an opportunity to save money, do it better and so forth. So our always starting point with the customer is, we want to be in the private cloud, converging infrastructure, giving you the building blocks of what it takes to go from legacy to modern. And I would not suggest that converging infrastructure actually is a gone phenomenon. I would argue that- It's changing. It is evolving. It is changing and it's evolving to the point that right now we can actually do the next click, so to speak, of converging infrastructure, where you bring together even a tighter level of power efficiency, it's a tighter level of operations and therefore a higher level for the application. Our conversation starts with, what do you want to get done? Can we actually make sure that we can help you with a professional services engagement if you like to, a consultative approach and that drives typically either data center conversations, security conversations, and user bring your own devices conversations. That is what we are both training our sales force and also trying to efficient ourselves. Okay, so basically power efficiency, accelerated data access to the apps, automation and scale with software and virtualization and application performance. All of those are the ingredients for the customer to see a quantifiable value in transforming the legacy environments to the new ones, John. Yeah, awesome. Well, I think that's what we'll be drilling on this week. Thanks for coming on theCUBE. This modern discussion, very relevant. Customers want to know what that means. They obviously, they're not throwing out converging. It's just changing. They all have to buy servers, storage and networking. I mean, those are the native ingredients. It's just how that's changing with virtualization. I really appreciate your conversation here. Very relevant. This is theCUBE. We'll be right back with our next guest right after this short break. Thank you. The video news business believed the internet was a fact. The science has settled. We all know the internet is here to stay. Bubbles and busts come and go, but the industry deserves a news.