 Hi, my name is Stu Miniman and I'm a senior analyst with the Wikibon project. I'm going to talk today about how enterprise architectures are invading the enterprise and how that's affecting converged infrastructure and really how all of IT purchases their infrastructure. So if we look at models in the enterprise today, enterprise today is they will spend their dollars to save time as opposed to hyperscale players. We're talking about the hyperscale players, talking about companies like Google, like Apple, like Facebook and they have a much different scale than the traditional enterprise players do. They spend their time to save money. So, you know, you talk about some of these guys and you say they might have a thousand PhDs to figure out how they can architect something. If you look at what Facebook is doing, they can actually manage over 20,000 servers with a single person managing that environment which is just orders of magnitude better than what the enterprise can do. And what we're seeing is really, despite the fact that hyperscale is only about 10% of the spend in IT, they're really driving a lot of the innovation. There's hot technologies like server-based flash and some of the sweet spots for software-defined networking or SDN are deployed in these environments. And hyperscale, when they buy their infrastructure, they really look at these as being disposable. What we're talking about is they really want to automate things as much as possible and reduce how much their people are touching these pieces. And we're starting to see some of the best practices and technologies from the enterprise pushing, from the hyperscale pushing into the enterprise. But one of the big differences for the enterprises, since they don't have all of those PhDs, what they really need to have in the enterprise is solutions. And what I mean by solutions is that somebody needs to take all of these components, put it together and offer it to them. It's not just buying white box servers or some kind of bare switches and saying, go build your own because there are only so many Googles out there that can do that kind of technology. So when Wikibon looks at this space, this architectural change, what we think is going to happen is the deployment of what we call software-led infrastructure. And software-led infrastructure is following some of the things that we've been watching for the last couple of decades, but some nuance and some specific things that we can now do with the technologies that are available. The first thing we want to look at in this environment is that we're talking about dynamic infrastructure. And by that we mean scale out. So I should be able to buy small and be able to add pieces incrementally. We've seen some point storage solutions out there in the past that have done scale out really well. If you look at Equalogix, if you look at Isilon, all of those really can buy piecemeal and they just scale nice and simple. Secondly, automation. If you hear what VMware talks about in the Software-Defined Data Center, automation is key and we would agree that automation needs to happen. Software-led solutions and in many cases this software can manage all the services and it's more than just managing the environment, but really functionality that might have been baked and hard-coded to the hardware before can now be led by the software and should be in many cases hardware agnostic. Also, simplification is absolutely critical. Need to get rid of what we like to say is the undifferentiated heavy lifting. So if you look at what Converge Infrastructure's been trying to drive and what the hyperscale people do, it's what we call rack level convergence, which is rather than thinking about my compute and my storage and my network as distinct silos of operation or cylinders of excellence as they were often called in the enterprise, we really just want to think about what processing, what storage and what bandwidth I need and it should be able to be deployed in a simple manner. And the last thing we talk about is there's a whole lot of new architectural designs that need to be considered, web-scale applications take a different architectural approach to how they're done and there are a lot of new technologies specifically want to call out. Flash is something that is drastically changing how architectures are built and cloud. So, you know, cloud is much more than just virtualized, it really is an operational model that needs to change and as Wikibon CTO David Fleuer says, when you build your own infrastructure sometimes it might want to be in-house, sometimes it might want to be in the public cloud or more and more you might even want to be thinking, can I have my information within a close proximity to access more cloud solutions? So, if you look at what Amazon has with its direct-net technology and big hosting providers like the Switch, Supernap in Las Vegas where we can actually have my data and cloud data close to each other and we can even use technologies like Backhaul to be able to plug high bandwidth connections between them because if we talk about going over distance, you know, data takes a long time to transfer so we need to really think about that. So, when we talk about converged infrastructure, converged infrastructure really has always been about simplification and if you look at the architectures that built for converged infrastructure, most of them are driving towards this rack-level convergence. So, of course, VCE's VBlock is the early leader in this space, Cisco with both VCE as well as with NetApp with the FlexPod, EMC with the VSpecs have been driving that rack-level convergence as well as IBM with their relaunch of pure systems about a year ago as well as HP who really has done a phenomenal job of converging the server and the storage so that many times you don't really think about them as server or storage but it's just that resource that we talked about that go together. And Dell's also driving converged infrastructure solutions. Beyond simplification, most of the converged infrastructure vendors from the traditional side are looking to drive automation and that's usually some layer of software that is put on top and it's a management way of looking at things. There's also ways to look at really kind of pooling and automating what I do with my infrastructure and once again, sometimes that requires an additional layer of software or some kind of clustering to be able to pull things together where the kind of traditional converged infrastructure solutions that have been on the market have been falling down is first of all, from a scale-out architecture, most of them don't fit that model real well. I need to think of some kind of matrix or clustered configuration and if we look at things like Flash, most of these architectures were not built specifically with Flash in mind and therefore they've had to, while they can put Flash into the environment, they're not necessarily taking advantage of server-based Flash in the best manner possible and the interaction with cloud is often hazy. Now I want to talk for a second about some of the new vendors, what the big guys would call the ankle-butters of the world but others have called hyper-convergent, architectures built with really that hyperscale mentality in mind and pulling that into the enterprise. So the first one I'll mention is scale computing. Scale computing has really built for the SMB market a solution that pulls that hyperscale architecture into the environment. It starts at a real small price point and really starts below where Flash today can hit from an architectural standpoint. So it's a nice solution to look at for people that want the efficiencies and operational models of hyperscale but not ready for not driving the applications that would be kind of in the middle of the market or the enterprise market. Second one is simplicity. Simplivity is actually based in Westboro, Massachusetts which is close to our office here in Marlboro. And simplicity, really what Wikibon believes they've done is build almost a God box. It takes a lot of the storage efficiency technologies and drives it into that kind of hyperscale model but since they're trying to put so many different features inside of it they have multiple fronts that they need to fight against. The third one that I'll mention is Nutanix. So Nutanix actually does the best in Wikibon's opinion in meeting the criteria that we've set forth here. It's built with really kind of the Google type architecture from a scale-out perspective so that they've learned from what Google file system and Colossus have done to be able to scale out their environment. They are a software-led architecture. So while they do offer the hardware pieces it's all commodity-based hardware and they offer Flash in a couple of different modes including server-based Flash and they built their design so that they can be hardware agnostic and keep adding on top of it. It is rack-level convergence so that it gets all of your server, storage and network in one piece and as I said they are built for Flash and for Cloud. They've actually came up with a term called Virtual Computing Platform or VCP where they've kind of matched out how they feel they meet these kind of new architectures and even talk about how Nutanix can coexist with cloud architectures really well so that I can have applications that can span between my internal and external what they call a hybrid environment. So the action that we really talk about at Wikibon for enterprise CIOs that want to look at converged infrastructure and how Hyperscale is going to invade that market is what you can get by moving towards converged infrastructure. Of course simplicity is one of the things that all of the vendors will tell you that they're driving towards and you need to look at not only what the initial acquisition cost is but what ongoing support is going to cost you. So you want a solution that is going to be able to help drive down your initial cost and secondly you want those operational efficiencies so productivity should go up when you're buying these solutions so look for environments that can really automate your environment and dynamically scale out so that we don't have to have our people spending so much time adding and tweaking and adjusting the environments that I have and finally the solution should also bring your workforce agility so we should be able to be doing more being able to spin up more applications and not have to think about okay when I add a new application what's the bespoke environment that that's going to go in and how long will it take me to tweak that environment you want to be more like what Facebook can do which is just okay I've got five different models to choose from I go choose the one and I go move forward on it so Wikibon really believes that we can look at the hyperscale solutions out there and have that be a predictor as to where the enterprise is going to come over the next couple of years and converged infrastructure and especially those hyperscale type solutions are going to help enterprises take advantage of the next generation of IT so this is Stu Miniman with Wikibon check out wikibon.org for all the latest research on hyperscale converged infrastructure, big data and much more and find all the videos with our media partner SiliconANGLE at youtube.com slash SiliconANGLE thank you for joining us