 Okay, we're back. This is Dave Vellante and I'm with Stu Miniman. We're with Wikibon.org and this is Silicon Angles theCUBE, where we bring you the smartest people that we can find, we travel the world, we go to events, we drop in, we extract the signal from the noise and we love to talk to customers, to get the customer angle, the practitioners who are in the front lines, fighting day to day, bombs are dropping, they have problems all around them and they're fixing them and it's, we love to cut through the messaging, the messaging great, we love to have the executives on but it's really the customers that are making it happen. And we've heard, we're at the Dell Storage Forum, we've heard a very strong message around the customer first and that's really always been Dell's heritage, my course, the legend, Michael Dell, started his company as dorm room and very customer focused. And we're here with Pete Keeler, who's a senior system admin at Techplot, a Seattle based firm. Pete, welcome to theCUBE. Oh, well thank you. Thank you for having me. Great to have you. Good to see you at this event, the second Dell Storage Forum. Were you in Orlando last year? Yes, I was in Orlando last year. Yeah, so I mean, it seems to be a step up, it's getting bigger and better. I mean, what do you think from a customer's perspective? It is fantastic. First of all, I have to say that I had a great experience at the DSF in Orlando and wanted to make sure that I returned for this event here today. So, having a good time so far. Yeah, good. Tell us about Techplot, what's the company all about? Techplot is a software company that makes simulation analytics software and data visualization. They make the software that really focuses on helping scientists and engineers solve really complex problems. And that has been their focus and they've been in business for about 31 years now, which is really unique in the software world. Well, so this big data trend, you guys must be loving that, right? Because squinting through all that data is really hard. Most certainly. So is it like Tableau? I'm just trying to understand. Yeah, it's similar. Many of our customers are in the aerospace industry and so they are trying to solve really difficult problems around how to make an airplane wing more efficient, how to fully understand some of the anomalies that might happen with aerodynamic efficiencies. And so this is really what our focus is on for our customers. So what are the big drivers in your business? I mean, obviously invention, innovation, what else? I mean, what are the drivers that are affecting your technology? Well, one of the challenges that we are faced with, of course, is this notion of big data. Our customers are seeing that with the software that we produce for them. And so our software is extremely sophisticated. We really needed a sophisticated infrastructure so that we could focus on developing our software in a faster, much more agile way. Okay, so talk about that infrastructure. Well, in late 2008, we had a very traditional infrastructure and I went to one of Adele's channel partners, Mosaic, at Technology, and went looking for better ways to help solve this issue that I had about manageability, especially for an IT department with very limited resources. Managing the data, the infrastructure? Managing the actual compute infrastructure and the storage services as well. And so I worked with them to come up with a solution on how could we have a scalable infrastructure that allowed us to meet the demands of storage and compute services. So can you be more specific? What was the real problem that you're having? Your business requirements are changing, you couldn't meet those requirements, the old alignment problem, is that right? Yeah, it was the classic case, especially with our software that we're working on. Especially with our software development group, that they needed a lot of computational power to have these build systems that would help compile the software that they were building. And being able to provision those systems in such a way that would benefit their practices was very difficult in a traditional infrastructure. Benefit from the standpoint of the speed to which they could. That doesn't VMware or Hyper-V solve that problem? That made a significant step and it helped us go in the right direction. Of course, we needed the actual physical infrastructure for VMware to live on. And one of the significant components that we had with respect to this new virtualization solution was how were we going to store the systems plus the data that the systems serve up? So, like a lot of customers, you virtualize your servers and then all of a sudden, it really impacts the storage. It's like a domino effect on the storage. You got IOs going crazy, you got less physical resources. So that's what was happening to you? Yes. You're a storage, we've heard a lot about legacy storage. We know that's a euphemism for old clarions. Is that what you had? No, we had, whether fortunate or unfortunate, we had not made any investment into shared storage at the time. Okay, so you direct-attached storage? Yes. Which is great, but it's not shared. Right. Okay, so you didn't even have a storage person at that point, really? No. You were just moving the infrastructure and so where did it go from there? Well, we really had to say, okay, if we were going to build out a new virtualized environment with shared storage with a limited IT resources as far as the actual manpower, what were we going to do, what solutions were available that would allow someone like myself to be able to dedicate the time to managing the storage side. And that's really where we made the decision to go with Adele, an equilogic solution for our shared storage. So you would, was it fair to say that you're an IT generalist? Yes. Okay, you're not a Lund Management guru? No, no. Congratulations. Yeah. I'm happy for you. Thank you. Okay, and so what was the experience like? So you go from this DAS environment in your VMware shop? Yes. Okay, so now you have VMware, that's all the bit, probably what, three or four or five years ago? That was, yeah, just shy of four years ago. So that created some great opportunities but also created some challenges, particularly in the storage side. So now you're bringing in equilogic, it's a new world for you. How did that go? What was it like? One of the most significant challenges that we had early on was simply acceptance of this new way of our storage and computing environment. Acceptance of the users? Yes, yes. So that was, these were strange new technologies that we were leveraging. And being able to help demonstrate how these technologies would benefit everyone in the organization, thus our customers was something that initially it took a little bit of work. What changed for the users when you brought in equilogic? The speed of provisioning significantly changed. Our software development group had automated build systems that would need me to provision them or to protect them in a very quick way. And that was simply not possible in the past. Why was there some friction initially? What was the concern? I think in just for the lack of being familiar with modern day storage technologies as well as computational services. So the notion, we have a very curious bunch there at Techplot and they simply want to know how things work. To say that we were going to abstract out computing services from storage and so forth was very foreign to them as well. Okay, so they just didn't want to change. They didn't want to change initially. Okay, and then take us through what happened. So you had to, you took some arrows. Trust me, it's going to work. I'll stick my neck out, you know. Okay, so your job's on the line, basically. You've got to keep your constituencies happy. What happened next? From that point on, it was really about being able to take advantage of the new tool sets that we had there and that we had available to us. From there, it was being able to provision these systems and to demonstrate how we could protect the systems in such a way that was going to be beneficial for not just our software development group, but the entire organization. At that time, it was about six months into our transition, we had a neighboring building that burned to the ground, which really helped the notion of being able to take these newer technologies and help leverage them for our disaster recovery initiatives. Okay, so the, can you be more specific? I mean, was that a VMware high availability capability combined with Equalogic or? No, it was a purely Equalogic Sand-to-Sand replication is really what I focused on. Replicated to off-site, a synchronous replication, or was it? Okay, so basically, what would have been the outcome had you not had that capability? Well, in the case, and let me be clear, this was a neighboring company that had this issue, so we were looking at that saying, how would we be able to recover from that similar scenario? Okay, so it didn't happen to you, but you went just at first hand, had it happened to you, it would have been a true disaster, right? Yes, yes. Okay, so your management, your team, they saw that neighboring company and they said, oh my gosh, if that ever happened to us, so they immediately knocked on your door and said, what would happen if that happened to us? And you said, no problem, we got it covered. Yes, it most certainly helped us accelerate our disaster recovery initiatives. That was the simplest way to make sure that that was going to be one of our efforts. Okay, so you were in sort of mid-process of setting that up, which is great because you didn't get snake bitten when your neighbors did, which is unfortunate for them, but you didn't have to absorb that, because a lot of times people closed the barn door after the horse leaves, so did they get promoted after that? No, but what we did do is we simply, yeah, learned the lessons from others there and it was a great way to see and to really demonstrate how quickly something could turn so wrong for an organization and we simply didn't want to be in that position. Since we made the investment into the ecological family of storage, many of those tools were built in at a very fundamental level, so we wanted to be able to really highlight those and make sure we were utilizing them. So how did backup change? Let me ask it that way as a result of this new initiative. For us, because we had a very traditional infrastructure, we, much like everyone else, were focused on serialized traditional backups to tape and there's only so many hours in the day and so many hours in a given week. Bowing down to the backup window. Yes, and that was one of the most significant challenges that we had, plus we run a very mixed environment there, so we were solely dependent on the different kinds of operating systems that we were trying to protect. When we were virtualizing our infrastructure and then moving it off onto shared storage such as the Ecologic Array, that removed that element of having concern over the specific operating systems and it took the focus on to block-based SAN replication. Okay, and so how do you do backups now? Backups I like to think of as far as an overall protection scheme there are localized snapshots that live on the different SAN arrays. There are off-site SAN to SAN replicas that live at a colo site and then there are long-term archiving methods as well. So are you providing backup as a service to application owners on a application-by-application basis or is it more of a one-size-fits-all? Since our infrastructure has really changed into virtual machines serving specific applications and being very focused on specific applications, I think it falls into the former there where it's per application. Okay, so you've got a form of IT as a service. Yeah. So a lot of customers that we talk to, this is changing but historically it's daily incrementals, weekly foals, that's it, take it or leave it. But you've got different RPO and RTO requirements. Do users even have a sense of RPO, RTO or do you take care of that? How does that whole business discussion occur? One of the things that I have learned in IT is that you have to be the advocate for these kinds of discussions and so I am always helping our management team better understand some of the implications about an organization such as a tech plot choosing to not address some of these issues and so it really becomes an education process that I provide our management team in helping them really understand what these things mean to them from a business perspective. Yeah, because I see people say RPO and RTO, those are business notions. We used to go to talk to a business person about RPO and RTO, they go, what are you talking about? I don't know what that means. How much data do I want to lose? I don't want to lose any data, what do you mean? So you're saying you translate the notions of RPO, RTO, data loss, et cetera into a conversation with a business person. So how do you do that? You sit down and say, okay, let's talk about your application, what would happen if you lost it? Do you have that, is that how the conversation goes? Many times yes and it's really not about what you can afford but what you can afford to lose in these scenarios where even if it's not necessarily a pure disaster, this monolithic sort of a disaster, but primary services being down, what are you willing to give up? And that is something that the requirements are built off of those discussions. So you have a pool, let's say, let me describe what I'm hearing. So you have this virtual pool of capability, compute, storage, and backup services that can accommodate your application portfolio and you allocate that pool according to the outcome of those business discussions and you assess capacity requirements on a fairly regular basis and you update that notion regularly periodically, is that right? Yeah, a lot of it simply comes from experience of understanding some of these critical systems and what sort of recovery periods we need in order to maintain normal business operations. But in concert with that is helping our management team understand what are simply the capabilities now. That's a conversation that has dramatically changed the very fact that I can not only protect the data but protect the systems that serve up the data is a profound difference as it was just five years ago. P, what do you make of Apashure? Is that something that you're looking at or do you not have to look at that because you've got your snapshots all set with the ecological rate? No, in fact, I was at the breakout session a little bit earlier this morning with regards to Apashure. So it's intriguing to you. Personally, I love the vision of being able to dial up or dial down an RPO based on the requirements of the app and it seems like a nice pickup for Dell. They have some very compelling features built into them. Okay, good. Last question, any advice that you would give to your peers that are looking at doing some type of similar initiatives? Yeah, I think one of the biggest things that has worked out well for me with regards to our efforts in virtualizing our infrastructure with shared storage is helping communicate the right message to the key stakeholders. Helping them understand, in very simple terms, it certainly doesn't need to be complex. What are the implications of moving forward versus not doing anything? And the more they understand that side of things, more than likely they will understand your position. I actually had one more question. You talked about one of your business drivers being, you know, agility. Are you guys moving toward a DevOps type of environment? You know, that whole notion of intersection of application development and infrastructure operations? That is certainly an intriguing element moving forward. I know that we have a software development group that is very strong into the DevOps side and the notion of bringing that into IT infrastructures is, again, certainly a compelling one. Yeah, I saw somebody in the elevator the other night from Munder Capital who we had a peer insight at Wikibon with Wolfgang Gerlich who runs the IT operations at Munder and they're a full DevOps shop and kind of exciting what people are doing, you know, creating hyper-productivity essentially with these new methodologies. So something that we've been watching, you can check that out on wikibon.org. Just Google Wikibon and Munder Capital and you'll find it. Okay, well Pete, thanks very much for coming to theCUBE. It was a pleasure meeting you and hearing the story of Tech Plot. Thank you. Keep it right there. This is the Dell Storage Forum. This is theCUBE. We're live from Boston. We'll be right back.