 Hi, welcome to CUBE Conversations. I'm Stu Miniman and we're at the Wikibon office in Marlboro, Massachusetts. Happy to have on the program first-time guest Chris Ferry, who's the co-founder and CTO of Store Magic. Chris, thank you for making the trip across the pond and welcome to Wikibon. Thank you, I'm glad to be here. All right, so Chris, about 10 years ago you were one of the founders of Store Magic. Can you bring us back as, you know, what was happening? What's your background and what led to the creation of Store Magic? Well, I'd been working in storage for probably about 15 years and for about five or six years in ice-guzzy storage. We thought we saw an opportunity for ice-guzzy storage, which 10 years ago was just beginning to take off. We were particularly focused on smaller businesses, the SMB, and producing sort of low-cost, easy-to-use solutions. Fairly quickly, we refocused away from the SMB to the distributed enterprise and to virtualized environments where we saw the advantages of bringing a simple, low-cost, highly available storage solution to the distributed enterprise. Yeah, wow. I think back 10 years ago I was working on ice-guzzy stuff too. You know, many looked at it as a disruptive alternative to some of the traditional, you know, kind of big block storage options that are out there. I mean, a lot has changed in the last 10 years, Chris. So, you know, when we think of Store Magic, you know, what's the focus of the company today? So, the focus of the company today is producing low-cost, highly available storage solutions for the edge in a virtualized environment. So, where you need applications at the edge which need to be highly available, but you don't want an expensive solution, you want something that can just run on a pair of servers using the internal storage in the servers, you don't want to buy external boxes or anything like that. So, our software runs as a virtual storage appliance on a hypervisor. It virtualizes the storage that's in the server and makes it highly available and then presents it to the environment so that if you get any failures, then it will instantly fail over to the other side and you've got complete high availability. Yes. So, in some ways, similar to what we've seen with virtualization, but the edge has a lot of different characteristics. I mean, one of the things I think, you know, most edge environments, you know, I might not have not only, it might not have a storage guy, I might not have a virtualization person there to manage IT. You probably don't want somebody managing your IT at the edge. What do you see in your customer environment? Exactly, yes. You know, a lot of the sites that we're in, it's a pair of servers, they might just be in a cupboard. There's no IT people on site or anything like that. What they want to do is you've got a, you know, a data center, a head office, they want to roll out a solution to the remote sites and then basically forget about it. They just want it to work. If something goes wrong, if you get a storage failure or something like that, then they want to know about that. So, you've got to provide all the right management tools so that you can monitor the environment, let them know when something goes wrong, and then just provide very simple ways of recovering from those failures. Yeah, so I think back, you know, 10 years ago, we talked about iSCSI. You've seen software-defined storage roll-in, discussions of hyperconvergence, you got Wikibon talking about server sands. Has the discussion changed a lot from the customer standpoint? How do they look at it? What kind of architectures do they build and, you know, what's the role of hardware in the solution that you guys put together? Well, I think the role of hardware has changed. I mean, a few years ago, sort of storage was all about hardware, about external boxes and often expensive external boxes. Now, the focus has changed. It's really about the software that's providing the storage services, which can run anywhere. It can run in a virtual storage appliance, and you can just use any commodity hardware you like. The other thing that's changed is go about a few years, and if you wanted to get to high performance, then you need to put loads of disks there, not for the capacity, but just to get the number of spindles up to get the performance. Now, when you throw flash and SSDs into the mix, you can get the performance without needing large numbers of disks. So you can put together some pretty impressive, highly high-performance systems with maybe just a small amount of flash storage and just a small number of spindles of conventional hard disks. All right, Chris, could you maybe talk a little bit about, you know, what's some of the key use cases you see? I know retail is one of the, you know, bigger customers where you've got, you know, deployment with thousands of, you know, remote sites. But, you know, what are some of the applications? We title this segment, you know, small data, big outcomes. You know, why is the data so important? Why is it critical to company businesses? Well, it obviously depends what the data is. So if you're talking about retail, then, excuse me, then typically at the store, they'll be running a number of applications, they'll be running pointer sales, stock control, all of that sort of thing. The data there obviously is critical and the systems can't go down. So, you know, the data needs to be protected and it typically some of the data will get sent back to the data center for, you know, for analysis or whatever. Other sites, other sorts of environments we're in, things like process control, wind farms, for example. These may be at very remote locations or maybe out in the middle of the North Sea or something like that. They've got a large number of windmills there. They all need to be controlled. They're gathering data from a lot of sensors. You know, this is really Internet of Things stuff. They're gathering data from a lot of sensors. They need to be monitoring and controlling the wind farm and they need to be sending some of that data back. But typically they'll be doing analysis at the edge and then sending back, analyze data back to the center. All right. You brought up the buzzword lately, IOT. Wikibon CTO David Floyer wrote an architectural piece talking about how all those edge devices, we're going to really have to, you know, manage process and handle that data more at the edge than back at a centralized location. What's your view of the world? How much data lives out at the edge versus, you know, pulling it back to some kind of central location to manage it? Well, the issue you've got with pulling it back to the central location is, you know, how big is your pipe? If you're at the edge, say on a wind farm, as the example I gave there or maybe out on an oil rig or something like that, then your connectivity to the data center, to the cloud, is limited. So you've got to do the analysis of the data at the edge and just send back refined data back to the center. All right. So, Chris, let's talk a little bit about some store magic partnership. So you talked about virtualization. You guys started with VMware but also work across Microsoft and your Linux base. So you can talk first about the virtualization environment, what you see, what your customers are using. Yeah, we are biggest, you know, the majority of our customers are still running on VMware, on ESX. And, you know, that's where we started. We provided one of the first virtual storage appliances to run on ESX. Hyper-V, we've seen sort of quite significant take up over the last sort of three years. I think with the release of server 2012, Hyper-V became a credible hypervisor. Before that it was maybe a little bit iffy. So we've certainly seen quite a bit of take up there. We are looking at other hypervisors as well. KVM is particularly interesting because we run on Linux. And I think if you want small footprint software running on sort of lightweight edge devices, then KVM could be quite important. Okay. What about from a server partner standpoint, you know, where are you engaging with customers? What do you see in the field? We're engaging with all of the main server manufacturers. So we're doing quite a lot with Cisco, with Lenovo. We've got sites out there running on Dell, HP servers and so forth. Okay. Can you maybe unpack for us a little bit the Cisco piece, you know, Cisco's done a lot with kind of converged and virtualized environments that the UCS was designed for. What differentiates your solution in how you partner with Cisco versus the rest? One of the things is our solution is very lightweight. You know, the hardware requirements are very modest, but at the same time it's very scalable as well. Throw hardware at it and you get high performance. We're particularly doing work with Cisco on the UCS mini, which is an interesting platform. It can start off with just as few as two blades and then grow up to eight blades. We can run our software there. We can virtualize the storage that is actually on the blades, or we can work with external storage, which could come from a Cisco UCS C-Series, for example. We can basically provide the software to provide the highly available and scalable storage using Cisco hardware. Also, we've been doing a lot of integration with other Cisco management tools, particularly UCS Director. So if you want to completely automate your environment and automate provisioning, orchestration and all that sort of thing, then we're fully integrated with UCS Director now. So you can provision storage from there. You can provision the storage appliances out to the blades, whatever. All right. Chris, I want to give you the last word. Your customers, can you walk me through what a typical deployment looks like? Who manages those remote environments and how much do they need to do upfront? And then once it's installed, how much do they need to do after the fact? A typical customer, it could be literally thousands of sites. It'll all be managed from a data center. The management may even be outsourced to a company possibly in another country. What they need to do is they need to initially roll out the solution to lots of sites. One area we've done a lot of work is tools to actually simplify that process. So tools using PowerShell, so you can script deployment. So, for example, one customer, they were rolling out over 2,000 stores and they were able to roll out over 30 stores a night using this scripted deployment. It was basically a matter of create a recipe, then run the scripts and they're up and running it at 30 stores the next day. Once the systems are up and running, it's just about monitoring it. So we integrate with a lot of standard monitoring tools like System Center Operations Manager, VMware vCenter, with SNMP, whatever, so you can monitor the environment. And if something goes wrong, typically it's a storage failure, it wouldn't be a software failure, but if something goes wrong, a storage failure, then they need to know about that at the data center so they can basically send out a replacement server. And another bit of technology we've got is something that simplifies the replacement of a server so they can just simply ship out a new server and then the software will automatically rebuild itself to the configuration it was with the previous server which had failed. All right, well Chris Ferry with StoreMagic, really appreciate you coming in. So many discussions of how software changed the world, something you guys have been working on a decade and the value of data at the edge, something we expect to see lots more through 2016 beyond. So thanks so much for joining us. Be sure to check out wikibon.com for all the research on this area and siliconangle.tv for upcoming reviews. Thanks so much for watching.