 Okay, we're back live here at HP Discover 2012. I'm John Furrier, the founder of SiliconANGLE.com and I'm joining my co-host. I'm Dave Vellante of Wikibon.org and we're here with Henry Faster, who's a chief technologist at SHI. Now, some of you may know SHI as a software house international, sort of the legacy name, but Henry, welcome to theCUBE. It's a pleasure to be here. So, HP Discover, you guys are a huge HP partner. Tell us a little bit about SHI for the folks that aren't familiar with you. So, SHI did a business for 23 years. It's a privately held company. We'll be at $4.2 billion of revenue last year. We should be at about $5 billion for this year. Started, as you mentioned, as software house international, primarily as a software reseller, but over the years, developed into a full service bar of software products, hardware products, professional services and then as of last year, cloud services. Yeah, so that's amazing growth for a reseller organization. You do a lot of systems integration, I presume. We do. A lot of customer services. So, cloud is the big thing. I think it's a hugely disruptive trend, especially for the middleman. But you guys sound like you're embracing it. Talk about what you're doing in the cloud and how you're helping your customers. So, we looked at the cloud as a logical extension of the products and services we offer customers. And so, obviously, we sell software products, hardware products, professional services, integration services. Well, adding cloud services to that portfolio is just a logical extension of that. So, whether I sell someone a physical server or a virtual server, it's the same customer. Our customers are our traditional IT customers. Our cloud was designed around that concept. It's not a retail cloud. It's not a credit card swipe model. Financial cloud. Our cloud is designed to run production applications for our existing IT customers. So, there's been a big transformation in the every seller channel, system integrator channel, VAR, Vabs, whatever you want to call them these days. But, you know, the business has changed from box-moving to solutions. And everyone now is cloud-washing. Oh, we got to have some cloud for you. It's a great way to get in the door, right? For an integrator or a channel partner to push cloud. And some are doing good jobs, some are just starting. Tell us through what you guys done. Because in talking off camera, you guys have been extensive work involved in building your cloud solution. Because there's a lot of work involved. So, just take us through how long you guys been working on your cloud solution, who you looked at, and why you decided what you went with. Okay. So, we started, we made the decision to enter the cloud over two years ago. And started our development effort around what we refer to as our industrial grade cloud. We launched it at VMworld last September. Been in full operation since that period of time. We have approaching 50 man years of design effort into the cloud. We've taken a fundamentally different approach. As I mentioned, we targeted our traditional IT customers. And in order to do that, we had to change some of the design tenets that people think about for cloud. And so, one of the things that you typically see in a cloud implementation is that the virtual machines, we offer infrastructures as services, our primary cloud offering. The virtual machines, as they're provisioned, are provisioned into the cloud provider's IP name space. And so, many of the issues you hear about cloud adoption. Difficulty in using application migration, data migration. Those difficulties occur because the virtual machines exist in the cloud provider's IP name space. So, we turned that model literally on its head. Well, we created a technology that allows us to project virtual machines onto a segment of the customer's network in the customer's IP name space. And so, those machines are, for all practical purposes, identical to the customer's local virtual machines or physical virtual machines. So, we needed to choose a wide variety of technologies to do that, obviously. Hewlett Packard is a very large provider to us of technologies. Clearly, our dominant technology provider, but not because we started out to be that way. We ran a series of exhaustive RFPs around all of the technical components of this, and Hewlett Packard just ended up winning the lion's share of them. What's the challenge on the private cloud? Because we kicked off our CUBE franchise at EMC World 2010 when EMC was, before they had their great marketing campaign, they have now CloudMeans Big Data, they had Journey to the Private Cloud. And so, Private Cloud's been all the rage. Public Cloud has been popular, obviously, for developers and non-critical things, but you guys have cracked the code on Private Cloud with customers. Take us through what you do to harden that solution. What are the key milestones there? So, there are several approaches. So, we offer our cloud really in two very distinct ways. We offer what we refer to as our multi-tenant cloud, which runs out of our cloud centers. And then, we also offer a service that we call Managed Private Cloud. So, our architecture is based on a service core model. And so, a service core is a finite collection of server storage and switching elements that we replicate over and over again to add capacity to our centers. But we also have the ability to place those, those service cores, we call them vCores from a marketing point of view, to place those vCores in the customer's center. We still own them, we still charge the customer just based on utilization. It's a cloud appliance. Yes, literally, it's a cloud appliance. And so, customers can have exactly the same experience using either our multi-tenant cloud or our Managed Private Cloud in their own centers. And so, from a hardening point of view, our cloud centers are designed to have no single points of failure, so that the technical infrastructure is totally redundant. The physical infrastructure of our cloud centers is all tier three plus. So, they're dual-gridded electrically, dual UPS capabilities, dual generators. We attach to two separate sonnet rings. Each sonnet ring is brought in through two different D-marks, so on and so on. In the security space, all data inside of our centers, inside of our cloud centers, is encrypted. So, all data in transit and all data at rest. And so, a customer has a wide variety of very hardened choices either from our center directly or by placing our vCores in their center, so that their data never leaves their four walls. Okay, so you offer both an on-premise private cloud, cloud-in-the-box, if you will, and the customers can directly access what you would call your public cloud. And then, you architect a hybrid of that using a homogeneous infrastructure. Exactly. Which, you know, I think homogeneity is your friend. Absolutely. Business, especially the private cloud, I mean, or the public cloud, or the hybrid cloud, sorry, we did a survey last year and it was notable in that very few of our practitioners in the Wikibon audience were actually doing hybrid cloud. Right. They were very concerned about it. This year that number's way, way up. Are you seeing that in your? Absolutely. And it's in several ways. Obviously, our offering itself is intrinsically hybrid. You can have a combination of managed private cloud on your site and use, but also get virtual machines from our multi-tenant centers. You use the same portal to do that. It's transparent in that regard. But in addition, obviously, people need to move virtual machine loans in and out of different clouds. So one of the things that we did very early on was design a complete OVF management capability as part of our cloud. And so the open virtualization framework, from our point of view, is absolutely the best way to move standardized virtual machine loads in and out of multiple clouds. And so our intention has always been to make it not only easy to move into our cloud, everybody tries to do that, but make it just as easy to move loads out of our cloud as customers need to do that for whatever the reasons may be. So Henry, I wonder if you could talk about the changes that are occurring in your data center design and your network infrastructure as a result of all this cloud activity. So our preliminary design has held up extremely well. So we've had to make no changes of any consequence in our networking design, or our service core design that I described. And so for us, it's been primarily around expansion. So we're opening two new cloud centers this year to handle capacity. So I think from a learning experience point of view, most of the things that we've done are around high performance computing. So SHI was selected as the infrastructure as a service provider for the internet to consortium. We partnered with Hewlett-Packer to do that. And as you might imagine, a lot of the universities are very interested in high performance computing. And so our cloud was obviously designed for typical commercial loads, and universities certainly have typical commercial loads. But they were very interested in the opportunity to run a high performance computing, or HPC, in the cloud. So we've been working with the several universities, with Utah, with Indiana, Penn State, Notre Dame, to create a version of our cloud, both at the portal level and the vCore level, that directly supports high performance computing. Well, I was going to ask for the use cases that you've seen. HPC and the cloud is compelling because you can run stuff on demand, which is a benefit, basically have it by the drink versus the cat-backs investment. Outside of that, what are the use cases you're seeing from an adoption standpoint? That's always a confusing area for other folks that are taking that journey. What are you seeing that you can share with the folks in terms of use cases? Well, we've seen, again, because our cloud was designed to run production applications, see we see a very broad range of applications coming to our cloud. So probably one of the most interesting ones actually occurred yesterday. There was a transit of Venus past the sun that occurs once every 117 years or something to that effect. And so Columbus State and Ohio was responsible for hosting that for the astronomical community. And so they chose SHI Cloud to host that event. And so all of the web serving associated with millions of users watching the Venus transit of the sun was running in the SHI Cloud. So we see everything from something really out at that end. Well, that's an on-demand opportunity. You want to run exchange. So you see that full range of applications. So you've got, that's a great example of an on-demand spot resource. You need to have that. Then you want to provision that pretty quickly. What about retail? Can you guys do any retail? What's the status of the retail market? So retail is a very fertile area for us, particularly in the burst computing space. There is hardly any industry that is impacted by specific days of the year more than retail. So the retail industry is very interested in our ability to provide them burst capacity associated with that. And another interesting use case is related to that. It's the preparation for those periods of time. And so we have a number of customers who now use our cloud to launch massive numbers of transactions at their infrastructure in preparation for the Christmas season or Mother's Day or Valentine, whatever it might be. So that it used to have to stand up a lot of infrastructure to do that testing for a very brief period of time. So we have one customer who uses over 2,000 eight-way virtual machines to launch transactions at their own environment. And so retail is very interested in exactly that type of testing. Yeah, I mean, one of, I think of Tom, I had actually several practitioners in Wikibon, one in particular, they have to plan so far ahead when they make a new infrastructure move and then they got to freeze everything. So you're saying that you can really take that pain away. So it all falls on you. Yeah, so they stand up their test infrastructure, their transaction infrastructure. And then, you know, since it's cloud-based, you know, they shut it off when they don't need it, doesn't cost them anything, and when they need to run those tests, they spin it all up. Henry, we have one, we're getting the time hook here, but I want to ask you one final question. Just for the folks out there who aren't in the details, like we were going down at home on multi-tenancy and all that stuff, what can you share with them about the state of cloud, reality of cloud being? I mean, it's been hype over the years, over the past two years in particular, people have been doing a lot of, having a lot of success with the cloud. What would you share with them that you've seen the top three things that make cloud a reality? What technologies? Just any anecdotal perspectives? Well, certainly all of the key technologies, virtualization technologies, portal technologies, network technologies, you know, have not only continued to improve in their normal cycle, but have become optimized for the cloud. So there's a lot of work that's being done to optimize specific products for the cloud. So that helps us enormously. Anything I don't have to create out of SHI labs that I can use whole cloth from a vendor is enormous improvement. And you've seen a lot of that over the last couple of years. So the technology itself is more applicable to the cloud. Can you comment on security? Yeah, security, it's the same type of effect. Security companies are becoming very focused on cloud security and what is necessary to do that. There's even a subset of that that has sprung up, a security monitoring firms. So we use a company called Solutionary that monitors our security infrastructure and our switching infrastructure 24 by seven. So in the event of any type of intrusion, they not only notify us, but they're required to notify our entire customer base. And so a few years ago, you wouldn't even find those type of companies. So what you see is the traditional companies and new companies who are either creating or optimizing products civilly for the cloud. And the fact that that homogeneity that we talked about before, you can define a security incident the same way. You can import it on the same way as opposed to trying to manage that all on your own in the bulk of the cloud. Anyway, I know we're out of time, John. Thanks very much. It was a great session. Oh, my pleasure, really appreciate the conversation. Okay, we'll be right back after this quick break with our next guest right in this quick second.