 The information technology industry has been plagued by complexity for decades, particularly IT infrastructure. As organizations and startups and established companies try to solve problems and fill what people call white space, naturally people bring in things like more appliances and other capabilities that need to be managed. Converged infrastructure is attacking that problem. The problem specifically being that 70% of the spend on IT infrastructure for storage and servers and related equipment goes to people costs. So unless you're in the IT business, unless you're a managed service provider, you probably don't want to over-invest in that capability. Hi everybody, this is Dave Vellante at Wikibon.org and this is the Cube SiliconANGLE Wikibon's continuous coverage of what's happening in the tech industry, what's happening in infrastructure, software-led infrastructure, cloud, big data and the like. I'm here with Stu Miniman, my co-host today. And Doran Kempel is here. He's the founder and chairman and CEO of a company called SimpliVity, which is solving this problem with Converged Infrastructure and a new approach. He's a tech athlete. Doran, welcome to the Cube. Thank you very much. Thank you for having me, Stu. So tell me why right off the bat, why did you start SimpliVity? What was the thinking behind and the motivation behind SimpliVity? Well, as you said earlier, IT is complex. And over the last 20 years, what we've seen is a lot of the problems are solved. They're solved by virtue of adding yet another appliance, another new product that addresses a specific problem, a specific area of functionality. So what's the problem? The problem is that if you look at the IT infrastructure stack today, what you'll find is about 8 to 12 different products that are all purchased from different vendors. Let's go over the list. A bunch of servers running, say, VMware. Then a storage switch, then highly available shared storage. Show the graphic. Then potentially a data deduplication for backup appliance. Then potentially a WAN optimization appliance, focused on the left part of the slide. And then potentially a cloud gateway that allows your IT administrator to send data to some public cloud. Let's say Amazon EC2. And then you might have an SSD array in order to accelerate some applications. And maybe an SSD caching array in order to accelerate some other applications. And what's not seen on this particular graphic are a bunch of software applications. Maybe software applications in order to protect some VMs. Maybe a backup application, maybe tape. So what's the problem with that? The problem is that the IT team is purchasing 8 to 12 different products from as many vendors. And then trains as many professionals. And then manages this stack from almost as many panes of glass. So this is very expensive. It's very complicated. When I say that it's expensive, as you mentioned earlier, Dave, it's very expensive in terms of all the hours of expertise that are invested in order to learn and then operate this scheme. And there is a lot of hardware to be purchased. A lot of hardware to be tuned. A lot of space and power. So we think that all of that is complicated. And what we have done, but your question was why? So this addresses the why. And the what is that we've developed a technology that we call OmniStack. And a product that is called OmniCube. And basically- Love the name by the way. Cube. Thank you very much. Yes, Cube is a good name. Love cubes. And basically a single OmniCube includes on a simulated basis. We like that term all the functionality of all the products that I just mentioned. So basically, all of that functionality is part of a single, all-inclusive stack that runs on an OmniCube. An OmniCube is a standard x86-based system. So now that we have basically assimilated this vertical stack, what's left? Well, we can now have tens, hundreds, and thousands of those cubes anywhere in the world. They all are a part of what we call a global federation, and they're all managed by one person. So basically, a complete simplification of IT, a dramatic reduction of TCO in terms of the operating expenses and capital expenses. So you're an infrastructure software company, essentially. Basically, a lot of the IT, most of the IT, or the IP, excuse me, is software. We do have what we call an OmniStack Accelerator, and that's our own PCAE card with an FPGA that we use in order to add a lot of the value. And that is something that I haven't yet addressed, but what we do that is very unique has to do with what we call data architecture. So we've created a data architecture that is novel, and I would describe it as follows. Basically, the applications write the data. When the applications write the data, let's give a small example, they write it in very small increments. Let's call them data sets. And these data sets are 4 or 8k in size. What most of the storage companies have been doing, which we do not do, is they put these very small data sets in very large containers. And some of them put them in 15 meg containers or one meg container. So think about the container as this whole room. So now you have all these original data sets in these very large, slow, inefficient containers. What we have done is we maintain the data in its original 4 or 8k data size. And what we have done is we've deduplicated, compressed, and optimized all the data at inception, as is created by the applications. You now have these billions of small, very agile data elements with global namespace. So wherever the data is, we have access to the data. And what that gives the customers, or IT, are great benefits. First of all, since everything is decompressed and optimized at inception, it means that all the layers within an OmniQ, the DRAM, the Flash SSD, and the HDD are decompressed and optimized. So we reduce the amount of IOPS that are very expensive today. Secondarily and very importantly, now mobilizing data from Charlotte to New York to San Francisco for protection for data migration, et cetera, becomes very efficient, reduces the total amount of bandwidth that is required. And thirdly, and obviously, we reduce the total amount of capacity that a customer needs to store physical capacity on-site, thereby reducing the total amount of space, the total amount of electricity that's required, et cetera. So your question was, you're basically a software company, and I would say, yes, 80% of RIP is software, but we have this very unique accelerator that allows to generate a new data architecture that we think is optimized for the 21st century. And you can optimize it because of those small data sets. Exactly. You can place them where they need to be, and your system is intelligent enough to do that. Exactly. Okay. Now you also said during the, you said dramatic savings and TCO I think is how you put it, but what's dramatic? What are you seeing from customers? So basically one of the issues that we're very happy about having started selling the product earlier this year is that there's very wide adoption at the small end of the market or the low end of the market, all the way to some of the largest service providers globally. So if you think about manufacturing companies, WASAO Coated is an example of a Wisconsin-based manufacturer, and what they have done is they have purchased four OmniCubes, two in their main data center, additional two in a remote data center, and they're running all of their IT on four OmniCubes. So think about that. They have the functionality, the performance of a Fortune 100 company, and they're running on x86 commodity servers that we present as OmniCubes. So this is very powerful. One person manages all this environment. Now, step from this mid-sized customer to some of the largest service providers. We have a customer who's planning on building a data center with hundreds and thousands of nodes. And with our global federation, with this massive scale out infrastructure, they can use very inexpensive hardware, basically standard, or you could say commodity servers, x86, with our OmniStack, which is our technology, under whatever virtualization that they choose, it could be VMware, it could be KVM, which we're working on, and basically they can compete with Amazon. And I dare say that we deliver to them the most efficient IT infrastructure available in the market today. We have the most efficient data architecture. Think about it. All the data is due to compression optimized. Nobody does it, nobody talks about it. It is deemed impossible. And having come from a company, Diligent Technology, which I founded and which I sold to IBM in 2008, I know a thing or two about deduplication. It is very complicated. And what we've done is completely novel. So you've taken actually a lot of, you know, people like to talk about feature products. You've taken a lot of essentially what are feature products and placed them into a single architecture. Right. And people like to talk about software led infrastructure or software defined infrastructure and what it means to me and what it should mean to the customers is the ability to do what we think that Amazon and Google have done, which is basically to develop a very rich functionality stack, run it on commodity x86 resources under some virtualization stack. And we have developed and have delivered this rich functionality stack that runs on x86 and runs under your choice of virtualization. Today VMware, tomorrow, other virtualization stacks. So Doran, I think if you look at how infrastructure was built before, so many of those pieces were as what Dave said, kind of bolt on, add on. We looked it back up and archiving and replication and there were two challenges. One is not everybody needed those, but when they then bolted them on how they integrated together, you know, weren't necessarily seamless. So I can see that from SimpliVity's standpoint, having all those features means they work really well together and you're really well optimized. But how do you balance that kind of the intellectual property and expense of building all that code versus, you know, how many of our customers are actually using this today? That's a great question. And when we started sharing the vision of SimpliVity with customers, the response that we would got from people is impossible. They would typically say BS. But that was a good response. They said, if you can actually deliver that, I'll try it tomorrow. Because some of the functionality and the capabilities that we bring is functionality that you need three or four or five or eight different startups to deliver. And I think that the way that we've been able to encompass all of that and of course a lot of hard work and it took us three and a half years, three and a half years to develop this OmniStack. A typical startup typically delivers the product within 18 months, otherwise they're not going to get funded by the type of VCs that founded SimpliVity and that includes Kleiner Perkins, Accel and Charles River Ventures. But we have a significant vision and by solving the core problem, the problem of antiquated data architecture. Because if you're dealing with data sets that are virtually the size of this room as opposed to the size of this cap, it's not efficient. The whole notion of tiering has to do with putting the right data on SSD. And if you make decisions about the right data when you think about something in the context of this room, well, how efficient can you be? You either put a lot of cold data on hot media or vice versa. So basically we resolved that problem and we architected all the functionality that you just referenced around this data architecture, around three technologies that are basically at the heart of OmniStack. Number one that I already referenced is the data virtualization engine. This is the new data architecture with the accelerator that allows us to create these very small global data elements with global namespace. The other element that is perhaps tied to the notion of hyperconvergence is all the functionality that is weaved around the data architecture and our own file system. There was no way to build this high performance system. Each OmniCube delivers 25,000 IOPS and has 20 to 40 terabytes of storage. So this is a very powerful box. There was no way for us to deliver that by using some file systems that somebody else delivered. Once we have established that infrastructure, the SimpliVity OmniStack infrastructure, it's very easy to add more functionality because we're not building separate appliances. So that's the second technology, we call it virtual resource simulator. And that allows us to run on a balanced manner, a top single resource pool that are comprised of tens and hundreds of x86 resources. This allows us to have conversations and deploying very large environments. We're talking about hundreds of systems. The third technology that is pivotal and is part of that stack is what we call global federated architecture. And that has to do with the ability to manage all the OmniCubes or all the nodes anywhere in the world from a single pane of glass and that's vCenter within the context of VMware, which means that a single person through vCenter, single screen, sees all the data centers, all the OmniCubes, all the resident virtual machines, including the ones that we run on EC2, which is another very important functionality we'll talk about later, from one pane of glass. And by the way, this admin doesn't need to be trained no more than you and I when we step into a Hertz rental car. They just need to be familiarized with that car. Great, I wonder, can we talk about that global federation for a little bit? It's my understanding that most of your clients are doing multi-site and talk a little bit about how service providers, and you mentioned Amazon tie into the solution that you offer. Great, so when I think about our customers, I think of a few categories and the fact that there are a few categories of customers speak to the ubiquity and the elasticity of the technology. So as I mentioned earlier, we have customers who are local credit unions, manufacturing companies, and these people apply the OmniCube in what we call data center in the box scheme. Now, it's not a single OmniCube, they're typically using two OmniCubes for high availability. And as you said to all of them, have at least another OmniCube and another site for DR. Suddenly DR is something that anybody can afford. All you need to do is have an OmniCube in a colo or one of your other branch offices and you can use that as a target for your backup volumes. In addition, as part of the federation, we allow the customers to also backup and restore from EC2, which means that we run a virtual instance of the OmniCube, the OmniStack, if you will, on EC2. And by invoking the resources of Amazon, you have another member of the federation in Amazon. So basically the federation, back to your point, is any collection of OmniCubes anywhere in the world or on EC2. So when we talk about the smaller mid-sized customers, they typically would have two or three or four OmniCubes locally and a few others either in another site or colo, and that is their environment. If we go to the mid-market, those can be enterprises that might have many branch offices. This is a great use case. So if a customer has 10 or 30, and we are actually in dialogue with a company that has over 2,000 points of presence just in the U.S., all these people can benefit from small OmniCubes. That's not going to be the CN3000, which is something I'll talk about later, but rather smaller form factors of the OmniCube in those remote sites. And again, manage everything from one pane of glass. If you go further up or even stay at the same level, there are service providers. And what we offer is something very interesting to service providers in VARs. We tell them you have 100 or 1,000 customers. Build your own cloud presence. Build it on OmniCubes. You're going to be more efficient, and then I use names that I'm not going to mention here. You're going to be more efficient than Cloud Provider A, or you're going to be more efficient than Cloud Provider G, or Z, doesn't really matter. And we can't teach you to be as large as they are. So you can't buy hardware for the discounts that they receive. But we will give you a stack that is more modern, is more efficient than theirs, and has this new data architecture. So what those service providers, those VARs are doing, is they're establishing their own cloud presence, and they offer those capabilities to their own customers. So for instance, in that category, if you're a local Cloud Provider, or a VAR who has now become a Cloud Provider, you can deploy OmniCubes that you've purchased from SimpliVity at your customer site. They can run their applications on their OmniCubes, and they can DR to the OmniCubes that are part of your cloud. So I've now addressed, call it the mid to high end of the market. Then there is the mega market. These are the people who build clouds and want to compete with the prior Cloud providers that I didn't name. And what we offer them with OmniStack, because they don't need to buy their servers from us. So if you are a very large Cloud Provider, and you're willing to talk about thousands of systems, then you choose your version of an x86 commodity server. We will qualify OmniStack for that particular server, and you can now build your own environment, and your environment is going to look as follows. A layer of commodity x86 resources that we harness and render into a single shared resource pool, OmniStack that offers all the functionality that we've described. And on top of that, whatever virtualization stack you choose, does that address your question? Yeah, so this is important, because if you're a Cloud Service Provider and you want to compete with Amazon or Google or Azure, the last thing you want to do is try to compete head on an infrastructure service, you're going to get killed. So what you have to do is add value and differentiate. And there are a number of ways to do that. We've talked about that a lot in the Wikibon community, but the last thing you want to do is to worry about doing non-differentiated heavy lifting as a differentiator, because you'll lose that game. That's the race to zero. So what you're providing is the capability to allow service providers and customers to deploy resources in other areas where they can add value. So that's what I see. In my humble opinion, if we go back to the year 2000, when a lot of people started talking about, it wasn't called Cloud Services, but they were called Service Providers, they tried to be storage service providers by buying their gear from the storage vendors. Well, that doesn't work because it's too expensive. If you look at what the large cloud providers are doing, they have developed their own stack and they're running it on commodity x86 resources under their own version of virtualization. Software led. Yes, if you want to compete with them, you need to follow that model. Now, if you want to hire a lot of engineers and try and develop it yourself, that is a feat that very few players can afford. What we have done is we've developed the omni stack and now we empower you, Mr. Service Provider, to be as efficient as you can be within your size and leverage the fact that you might have proximity and relationships with a lot of those customers. Okay, you also have some news. So talk about the announcement that you're making. Okay, so clearly, when a company like SimpliVity on release one delivers all the functionality that I described, the skeptical question would be, what else can you now do? You've already delivered all the functionality and I admit that it is a challenge, but since we're speaking with small, medium, large and mega customers, they all require different form factors. And when we speak with different customers, sometimes the large ones say, I have 2,000 points of presence, I need smaller OmniCubes. Or when we speak with those same very large companies, they say, how about running my SAP on OmniCubes? Can you certify that? How about running Hadoop on OmniCubes? What about VDI? So we are being hammered, and I mean that very positively with requests for different form factors of OmniCube. The good news is that the technology is architected for that purpose, because all we need to do is take the same OmniStack that runs on the OmniCube CN3000, the same OmniStack that runs on EC2 and I'll deploy it and qualify it in smaller and larger form factors. So in that spirit, we are releasing today two additional form factors of OmniCube. The first one that we already delivered is a CN3000. It has either 12 and now 24 cores with about 20 to 40 terabytes of storage. We're releasing a smaller OmniCube that is called the CN2000 that has six to eight cores and has seven terabytes. So this is a smaller and less expensive OmniCube more suited for branch offices. And we're delivering what we call the CN5000 that is a juiced up OmniCube has twice as many drives. So 24 drives as opposed to 12 drives which are part of the CN3000 and more SSD and 24 cores. And this is for call it special missions. So basically, if you think about it when we went out earlier this year, there was one OmniCube. Now there are four different form factors. There are two within the CN3000, the 12 core, the 24 core and then you have the CN2000 and the CN5000. So that's aspect number one. Aspect number two is part of our strategy of technological ubiquity. We are offering, we've released a new form of the OmniStack or OmniCube accelerator. It's OmniStack when it's outside of the OmniCube, when it's inside the OmniCube, it's the OmniCube accelerator. Basically that's our PCIe card with the FPGA and we've delivered another one that is faster and is actually more compact. It actually looks like a bodybuilder a little bit but we are very proud of that child. In addition, customers have asked us to integrate with variety of functionalities. For example, VAAI of VMware was important to some customers we have delivered that and that is a category of new integration point and qualifications. And lastly, and this is interesting in the context of our architecture is what we call project 512 by 64. So what is 512 by 64? As I said earlier, we have developed on day one an architecture that is massively scalable. But in order for me to sell say 50 OmniCubes as part of Federation, we need to qualify that. So we have a qualification project that aligns with requests of customers that we already have according to which by this time next year we will have qualified 512 OmniCubes. That's about 20 petabytes of usable storage across 64 different sites, 512 by 64. Now there's nothing holy or magical about that number. It's just a target that we had to set. And by the way, it's part of a roadmap that goes well beyond the few hundreds into the thousands because we already are deploying in customer accounts that need more than that. So this is a very important project for us and it means that we're buying more hardware, more lab space, we just deployed eight more tons of cooling in our lab. So that's what the new release is about. More form factors, a more athletic accelerator card, more integration into different environments and the features and this project that we call 512 by 64. Excellent, Jordan, we have to wrap. But before we do, I wanted you to talk a little bit about the company. You mentioned some great VCs, Kleiner, Excel, CRV. So you've done a couple of rounds, I guess. Maybe you could talk about how much you've raised and also things like head count and then finally, what thing should we be watching as independent observers for the progress of SimpliVity? So we are launching our KVM program. If you're a KVM engineer, you want to give us a call. This is a very challenging, very interesting future project. The company is growing by leaps and bounds. We have a wonderful team and I'm really proud to be part of that team. Great investors will be raising money again probably at the end of this year. We've raised $43 million today but there's tremendous growth, tremendous opportunities and in spite of the fact that it seems as if we've solved all the problems, if you think about what we've developed, there's so much more that we can do in order to further simplify it. We have about three or four years worth of ideas that we need to deliver upon and we need the best engineers out there. And do you share head count? Can you tell us what you've got? We've crossed the three-digit figures about a month ago. Excellent. So we celebrated, we had a cake, the cake had the number 100 on it and it was not my age. That's great, nice critical pass there. All right, Dora, well listen, thanks very much for coming inside theCUBE and sharing this SimpliVity story. We'll be watching, we'll see you at VMworld, so look for that. Thank you for the opportunity. Thank you, Stu, for co-hosting. Thanks for watching everybody. This is theCUBE, SiliconANGLE and Wikibon. We'll see you next time.