 We're back, this is Dave Vellante with Stu Miniman, and this is theCUBE's Silicon Angles coverage of the VTUG Winter Warmer. theCUBE goes out, we extract the signal from the noise, and we're here at Gillette Stadium. Jeremy Wheeler is here, he's an innovation architect at Dairy Lee. Jeremy, welcome to theCUBE. Thanks for having me here. Thanks for coming on, and why don't you start with Dairy Lee, it's a marketing and services organization, but maybe tell us a little bit more about the company and your role. Sure, Dairy Lee is a milk marketing cooperative out of Syracuse, New York. It was formed in 1907, and originally it was just a way for farmers to work together to improve their milk price. Over the years, it's evolved. At one point, we had processing plants, then we got out of that business, and now we coordinate everything from the sale of the milk, the pickup of the milk, delivery from farm to the plant. We also offer health and workers' comp insurance, a lending agency, auction markets so that farmers can auction their cattle and their other products, and really, we try to be the one-stop shop, top to bottom for everything a farm would need. Right, so Jeremy, I'm guessing in 1907 that Dairy Lee didn't have an innovation architect. Can you tell us a little bit about what that is and your role within the company? Sure, my role is to identify technologies and evaluate the positive impact they could have on our business, and then also involves a lot of network architecture. So it's a little play on network architecture. Can you give us a little bit of a snapshot as to the size of your organization from an IT standpoint and what kind of gear you guys use? Sure, we have about 500 employees. We also service anywhere from 1,500 to 2,000 farms in various ways, and then our architecture, since 2005 or six, we started moving to VMware. We were 100%, or as close as we can get to 100% virtual on the servers by 2008, and really everything except phone systems or anything that would require some sort of a card or connection that can't be virtualized is virtualized. We, until recently, used HP Sands and we provide any kind of web services, SQL services, Exchange, SharePoint. If there's a piece of software out there, a major piece of software, we probably support it. So talk a little bit more about your business. So you've got this cooperative of 1,500 to 2,000 farmers in all or mostly in upstate New York, is that right? No, we span from Maine down to Virginia and over to Ohio, so we're all of the Northeast. Yeah, okay, and so logistics, you're providing shared services. Is there a marketing component as well that you guys do or sort of advocate for? We are the marketers, we'll, yeah, exactly. We coordinate all the purchase of the milk from the farms to plants, you know, Hersheycraft. You name it, if it has milk, we probably sell it. So what are your clients asking you for? What are the demands on your business? We have to do the milk testing. We work with another cooperative that does the testing. We provide the test results to the farms through a web service. That testing is what determines how they're paid on their milk based on bacteria, butterfat count, anything like that. And really the farms are all their own business and for that reason we have to be 24 seven because they get to decide their hours and we have to be there to support it when they want to work. Okay, so and tell us, paint up a picture. Stu asked you about the infrastructure. You talked about the virtualization. Maybe give us a little bit more insight as to what the infrastructure looks like. Sure, we're providing exchange, like I said, exchange, SQL, we're all virtual. We have nine massive ESX hosts and now we're using the SimpliVity OmniCube for our storage and for some of our process power and we do virtual desktop infrastructure. So we're doing VDI and servers and we're running around 500 VMs across those nine servers and on that storage. How much storage? Now we're running around 60 terabytes of storage and we're also replicating that to a disaster recovery site. So Jeremy, the project that you did for SimpliVity, my understanding is you didn't start out going looking for kind of some new server-based architecture. You didn't ask for hyperconverge or anything like that. It started looking for backup as my understanding. Can you walk us through what was the challenge you were facing and how did you come across SimpliVity and what was your decision criteria? Sure, the backup solution we had was problematic. Our backup administrator was spending a lot of his time just trying to keep the backups chugging along and he also had a full-time job of managing our tech support group. So really we were trying to alleviate the workload on him so he could focus on more beneficial things like the support side and the project started just to identify a backup solution and we had a parallel project to identify a DR solution and it was a six-month-long project and we had actually made a decision to go with an appliance-based backup that would replicate off-site but really it was just offering backup and it was just offering it for the servers, not for the desktops. We had made our decision but hadn't purchased and we went back to one of our vendors and said, okay, we need storage. We're starting to work on our DR project. And he said, well, you know, slow down a minute. I just saw something at VMworld. He used to check it out and he had us go through a SimpliVity demo and we were sold. It included the backup. It solved our DR issue in the same box. It has a major D-Doop on both the backup and the live storage and it provided compute power that we hadn't budgeted for and it was a refresh for our sands which wasn't even on our radar at the moment. So make sure I understand this. So you started with the backup project and you essentially ended up with an OmniCube which was doing obviously backup but you got some added benefit it sounds like. The compute, the converged infrastructure. A lot of added benefit. It was really for the same cost that it would have been to do the backup. Almost an identical cost just for the backup. And so you essentially recreated your entire infrastructure strategy. We did, we did. But that's kind of serendipitous. You start with the backup and end up with a new infrastructure strategy. It was a great find, it was a great find. And originally we hadn't intended to back up the virtual desktop infrastructure at all, just servers. That was a trade off that you had to make because. Because of the cost. But with the OmniCube, the amount of storage and with the way D-Doops were able to actually back up and offload the, replicate the VDI over to the hot site as well. What about RTO and RPO objectives? Maybe you could talk about that a little bit. That's obviously always a driver when you start with the backup. Where were you and where are you now? Going through the project, we went through each of our business units. Dairy Lee includes multiple business units and we got a sign off on each of them, accepting up to a 24 hour loss if there were a disaster situation. But you love those discussions. How much data do you want to lose? Yeah, how much can you? You're willing to lose. None! What do you want to pay for that? Oh, okay. Okay, so 24 hours was sort of the. That was the minimum that we had for many. The RPO. Yes. And with this solution, we're sticking to that for now. We're still early in it, but we're going to watch our data growth and assuming that the D-dub functions the way we think it will, we're going to go back and reevaluate those RTOs. We think we can do a much lower RTO maybe an hour. Oh, I'm sorry. That was RTO. Yeah. Oh, okay. What about RPO? Sorry. Yeah. So in other words, the RTO is the recovery time objective time. So within 24 hours you'll be back up online. Yes. Right? What about the amount of data that you, the business is willing to. Oh, I'm sorry. That is the RTO. Yeah, okay. From the 24 hours, we think we can go back through after we see how the D-dub is functioning and how we're doing on storage, we think we can get down to maybe hourly backups and get it down to where there's only an hour of loss. Yeah, because I was going to say 24 hour RPO for a business line. I mean, I wouldn't settle for that. I mean, come on. We can do better than that, right? So you're saying. I think we can get it down to an hour. And using some kind of snapshot, like. Yeah, yeah, the backup and the OmniCubes, we went from our biggest servers, one and a half terabytes, and it was taking, if we started the full backup on a Friday, we'd hope it was done by Monday. Now it's taking under 60 seconds. So we're able to do a full backup every hour if we choose and have that change replicated to the other side. And the software for that snapshot capability that. All included in the OmniCube. That's part of OmniCube. Yeah. Okay, so you get the backup infrastructure, you get the storage, the networking, the compute infrastructure and the backup software capability. Correct. All in the box. Yeah, yeah. Our recovery time objective is actually even lower than that. I mean, we can do a restore and same as the backups under 60 seconds, fire the VM backup, and we've done multiple test restores already in our DR site and it's incredibly fast. So Jeremy, I wonder if you can talk to us a little bit about the implementation. If you went in with kind of a, just backup in DR, which seems a very contained, understand what part of the organization you need to have that as opposed to a complete refresh. You've got migration from the old array to the new array. It seems like it would span, it was it a much bigger project involving more of the team or what was your experience? How long did it take? Kind of, can you put a rough, how many man hours you put into it? It didn't take long at all. The one piece we had to add is you need a 10 gig backbone for the OmniCube. So I went through and they'd ship me the equipment. I racked the servers and I racked the switches and I had all the wiring done when they got there. Maybe it took me 40 hours to get all that built and put together. And then Warren from SimpliVity came in, fired up the cubes and configured them in a day. And then the second day he was there was just training and us asking questions. So I was in a session earlier this morning listening to Chris Harney and Tony Asaro. They would tell you, I don't know if you were in that session or not, but they were talking about the cloud and the cloud as a potential threat to IT and IT roles. And I kind of wanted to have that discussion with you. So you've got essentially this converged infrastructure. I don't know if you'd consider it cloud-like or more services-oriented, I don't know what term you're comfortable with. I consider it private cloud. Okay, so you consider it a private cloud and probably building automation in as well. So do you feel like you're closing the gap with the public cloud service providers that you're able to add that type of value to your organization? That it is a reasonable strategy for you guys to continue there? Or what's your take on the whole public cloud? Is there a place for that in your organization? There is. The OmniCube ties into public cloud really well, especially in Amazon and those kind of services. So if we chose, we could add a third co-location site with Amazon and start replicating to there and make an Amazon DR site part of our federation. Yeah, so actually that brings up a good point. Yours is a two-site location. My experience is that adding kind of replication onto traditional storage arrays was almost an afterthought. One of the things that I found pretty interesting about SimpliVity is most of their customers are multi-site, you're the network guy, you own responsibility for that. Can you tell us how was that to kind of set up the two sites and walk us through a little bit of that? Yeah, the two sites were, we actually work with another cooperative and they do our lab testing and we came up with sort of a reciprocal agreement where they provide a space in their computer room and we provide them space in ours and we share the cost of 100 meg point to point connection and it was great that we had the site. We were having a hard time getting a good DR plan in place, we had some replication going on using SQL mirroring or DFS on our Windows file servers but there were huge gaps, especially Exchange and those kind of services. So this, since the backup and replication sort of happens at the hardware level and it does it all for us, now every server that we have on there is replicated to the other site and vice versa. So what do you think of this venue, Jeremy? We're here. That's great. My partner John Furrier is I'm sure really jealous but I'm sure he's plugging for a San Francisco Patriots Super Bowl, which I don't know about that second part of that, Stu. So we're here live at the VTUG, we're here at Gillette Stadium. Jeremy Wheeler is with Dairy Lee and unpacking a case study of how they're essentially sort of stumbled in, if I could say, to converged infrastructure. So going forward, do you feel like services that you can provide will be able, as you continue to automate, you've built that private cloud, it sounds like you'll over time make connections to the public cloud. Will this carry you through the next decade or how much visibility, how much runway do you think this approach gives you? Our plan was three years. We wanted to have enough storage that we could handle all of our backup in DR for three years and again with their D-dup and compression it looks like we're going to possibly go beyond that but if we find we need more storage, the great thing with the OmniCubes is we can buy another, plug it in and it just expands the available storage. It becomes part of the Federation and you expand it in place. It's not a major deal to add more. So I'm wondering if you could paint for us, we painted a little bit the before picture. You had the poor backup guy, kind of had his full-time job and backup was taking over that. You're the network guy and sounds like you were heavily involved in this. What did your team look like and how much time are you spending on the infrastructure today versus what you did before? My team is myself and then we have the tech support manager and he has three more tech support personnel under him and he was easily spending 75% of his time just trying to wrestle backups into working and our strategy after we put the OmniCubes in was to move the problem servers first, the ones that were having problems backing up. It was actually such a quick migration, we just kept going and within a couple weeks we had 90 servers that we had storage be motioned from our old storage to the OmniCube and we've eliminated, he still has servers in the old backup solution until we decide to move more. We're still cautious in that we wanna see how the data availability pans out before we move everything on there but his problems have gone away with the backup and he's actually able to focus full-time on managing his support group and handling support. What about Flash? Are you experimenting with Flash or some of your applications sort of at that point where they can benefit from Flash? I wonder if you could talk about that a little bit. We haven't really experimented with Flash. So you don't see that on the near-term horizon? Is it too expensive or maybe not an experience? There's some Flash baked into the solution that you've got, correct? There is and it's not something we can actually manage. It's actually handled by OmniCube on the back end and it looks at the things that are being accessed the most and tries to load them into Flash that they're more accessible but in our personal management we're not actually getting into trying to manage them. From an application development standpoint, you're not trying to explicitly exploit the Flash infrastructure of the OmniCube and that's because your performance is fine or do you see that, so when you virtualize your infrastructure it didn't affect performance in a negative way? I wonder if you could talk about that a bit. No, we didn't really have a performance issue prior to this what we had was a space issue. So we knew we were coming up to a point where we'd either have to buy a new sand or a really big new sand or replace the three that we had and the OmniCube actually did that for us. It replaced all of the sands that we currently had and the performance we saw no decrease in performance at all and we didn't see an increase but that's because we didn't really have a performance issue to begin with. So Jeremy, last question I have for you is you went through this project, the project itself seemed a little bit of a surprise so what lessons learned do you have? If you look back on it and said, boy, I learned this in the process or if I could have changed something or advice that you would give to your peers out there what would it be? It's hard to say the process is what it is and I'm glad it led me to SimpliVity because this was, we never expected a solution that would handle so many of our outstanding issues. We knew we were running short on storage but that wasn't even on the radar yet that wasn't a project and I guess, I wouldn't change the process we went through because I'm glad where we ended up but I would tell people, if you're looking at backup or DR or storage, sit through a demo with SimpliVity because probably for an equivalent cost to what you'd pay for any one of those you're going to get all three plus some compute power on top of it and it's just a really, really full featured product. Jeremy, how about some of the, I wonder if I could ask you from an IT practitioners perspective, we always talk about these big mega trends, cloud, mobile, social, big data. Obviously, cloud is something that we've talked about a little bit here. What about mobile? How is mobile affecting sort of what you do? We're big into mobile now, especially iPad and iPhone. We have 70 plus field staff plus their management group equals about 100 people and we didn't really have a consistent training platform for them before. It was kind of sink or swim when they came in so our VP of HR went through and created this, we call it a knowledge library and it's just a glossary of all the things that they should know and then references to documents that, and we're talking 10,000 pages of documents of government, FDA regulations, anything else they might need to know. We digitized it all, put it into a SharePoint site and then using existing iPad apps we're now able to distribute all of that live to all of our. So you don't even have to write the apps. No, we were able to use all store bought and so we've got a huge mobile workforce that we're managing the content on their iPads remotely and mobile is huge for us. As a practitioner, where do you get your information on some of these trends and do you use, for instance, do you use social media? I use LinkedIn for social media but I attend a lot of these type of events. I've attended VMworld five times. I've been to multiple Vmugs and that's where I go to. So it's peers, a lot of information from peers. Some LinkedIn, not Twitter, you don't get your information from Twitter? No, no, I never got into Twitter. Okay, cool. All right, Jeremy Willison, thanks very much for coming on theCUBE, sharing the case study of Dairy Lee, good luck going forward and it was really a pleasure having you. Oh, thank you. All right, keep it right there, everybody. We'll be right back. This is Dave Vellante with Stu Miniman. This is Silicon Angles theCUBE. We're here live at Gillette Stadium. This is the VTUG. Winter Warmer, we'll be right back. We're live mobile studio and we bring it to events and we say we extract.