Veeam Community Podcast

The Veeam Community Podcast is focused on giving the global virtualization community a new resource for connecting with industry experts, bloggers and peers, as well as for staying up to date on the latest industry news, developments and trends. Each 30-minute weekly episode of the Veeam Community Podcast will be available through RSS subscription and the Apple iTunes store. The podcast is hosted by Rick Vanover, a writer, blogger, VMware vExpert, VMware Certified Professional, Microsoft Certified Professional, Microsoft Certified IT Professional and a Microsoft Certified Systems Administrator. He currently works as Software Strategy Specialist at Veeam.

In this episode, host Rick Vanover (Twitter @RickVanover) welcomes Doug Carson (Twitter @Douglas_Carson) and Stan Brinkerhoff (Twitter @sirstan) to talk about time savings on virtualization platform topics such as clones and templates. Additional conversation points are raised about blade servers and their power savings as well as management capabilities.

Both Doug and Stan are in the hotseat for the "Three views from you" element of the podcast.

Rick: Hello, and welcome to the Veeam Community Podcast. I am your host Rick Vanover. This is Episode 5, "Saving Time with Virtualization." Here we go. Today our guests are Stan Brinkerhoff and Doug Carson. How are you doing guys?

Doug: I'm great, thanks.

Stan: Great, Rick. How are you doing today?

Rick: Hey, I'm great for a Monday. This is a Monday, and I had a busy weekend. I am just ready to get back into it. Thank goodness for Monday. I don't think people say that frequently, but I just need to get caught up and get moving with things.

Stan: Sure. I'm not sure where you're from, but we're up in Vermont here. We had a gnarly snowstorm all weekend. So I was out shoveling ice with a pickaxe and just back to basics. No technology there that can save you from the ice on the roof.

Rick: Now Doug, you are in . . . help me out. Where are you, Doug?

Doug: I'm in Glasgow, Scotland.

Rick: Gotcha. Your Skype profile threw me off, because it's East Kilbride. Then I'm like, well, the UK is what Skype says. What's the whole difference between Scotland and Britain and England and the UK? I can never get it straight.

Doug: The United Kingdom consists of several countries - England, Scotland, Northern Ireland, and Wales. I'm in Scotland right now, but I like to keep you guessing. I've got a picture of the Sydney Opera House and the UK. So, I like to keep you guessing where I'm from.

Rick: Yeah, that is weird. I didn't catch that. That is the Sydney Harbour there on your Skype picture. Stan's Skype picture is appropriate with a ski mask and a beanie cap on I think.

Doug: Just to keep it even more geeky, I managed to take that picture on a webcam on top of Sydney Harbour Bridge. It's just to keep it [inaudible 02:04].

Rick: That's cool guys. Thanks for being on the show. We're just going to talk about virtualization. It's all about saving time nowadays. In my own personal practice, it's so easy to provision a server, so easy . . . I work for Veeam, so we can make it easy to back up a server as well. What kind of time savings do you guys see or use in your own virtualization practice around making your day more efficient as far as the operational tactics of your environments? What kind of stuff do you have going on?

Doug: I'll take that I guess. Yeah, pretty much. Having the templates, deploy templates. The fact that if you're wanting to go to the experience of things like Enterprise Plus and deploy host profiles and network distributed vSwitches and stuff like that, just being about plug a host in and configure it, check I guess the profiles, and off you go. With ESXi now you can install it with 15 or 20 minutes. You've got another ESX host in there.

It's just phenomenal the difference between what the physical environment used to be like and now, to be honest, the thought of even deploying a physical server fills me with dread, because to think that it can take half a day for a physical server and 15 minutes to do an ESXi host and deploy within 15 minutes VMs. It's just far better, far better. I personally find myself with things like that is to try and make my life even easier with things like CLI and stuff like that and really try and integrate these things to try to better the cost savings. I don't know about yourself, Stan. Have you been seeing these kind of things as well?

Stan: Virtualization has made a huge difference, I think, in every part and every aspect of what we do in IT. In Vermont here, we've got a bunch of smaller businesses. Big businesses aren't here as much. So in my world, we deal with a lot of these businesses that have got one, two, three, four is a good number of physical servers, or even logical servers. Did I just break this?

Rick: No, you were going fine.

Stan: Okay, sorry. I shouldn't play with things.

Rick: No problem. So just start off with here in Vermont . . .

Stan: Okay. Yeah, so here in Vermont, we've got a huge group of just small startup businesses that do tons of great things. It's a really great environment for that. But what kind of comes with that you'll walk into a business that's just got one server or two servers. What virtualization has really brought us here is just huge, huge impact on backup and recoverability and being able to test your backup. I know there are some products out there that can do instant checks of these backup files.

So you walk into a place . . . before these snowstorms . . . in Vermont we lose power every now and then. We lose our connectivity back to the Internet. But you can have a good test of your backup on a Friday before you go home, which is just huge. Back in the days with tape, your backup window is eight hours. After that eight hours, you've got to recover to a host. It was such a huge time sink that no one ever did it, and it's huge to know that you've got a good backup, and it's a huge timesaver.

Rick: Stan, you're doing me a favor. I try to sneak in my cheap plug for Veeam on every podcast, so you set me up great. So I saw a tweet of yours, and I'm just going to quote your tweet here, "Using Veeam Instant Recovery, once again, saved myself three hours of work. This purchase is about to pay for itself in saved time." That was actually the theme for the podcast, saving time. So you use Instant Recovery to do things like that and save time, so that's another way you can take a functionality and just use it for your larger time management and your virtualization environment practice. I think that's pretty cool that you're using it like that.

Stan: Yeah, I'm not sure who over there wrote that NFS server that's built into Veeam Backup and Recovery, but what a cool addition to the product. When someone comes to you and they're like, "Hey, I need this one file back or I need part of an SQL server recovered," we used to do just that one piece, and it would have been an investment in time that's been committed. If you miss a piece of it or you don't go wide enough in your backups, then you've got to start all over again. The ability just to recover instantly, it's almost faster than instant.

Rick: Oh, come on. You're being too nice.

Stan: No, it's like you right click on your domain controller and your exchange server, and 15 seconds later it's booting up. It's intense. How is that possible?

Rick: Well, it's officially magic is what we have going on.

Stan: I believe it. It's one of those little things you run into in IT. It makes sense. I have backups. I get that you can make a custom manifest server. I get that VMware can talk to that, and I understand all that. But it's like, when it's there in the moment, and it's just saved you globs of time, it's a wonderful thing.

Rick: I think virtualization just as a platform has done so many other things. Doug mentioned templates and deploying to server. The last thing I'd want to do is wait six weeks for the shipment of a server or whatever just for an application. That scares the bejeezles out of me.

Doug: Yeah, I don't know if you mentioned about Instant Recovery and testing backups and stuff like that automatically. Have you guys seen yourselves there's more and more pressure, even on the small businesses, to do these things, to be able to purchase two orders and things like that, the ability to ensure that they've tested these things, ensure that they've done restores. We are seeing that increasingly more with some of the business clients I've talked to.

Stan: What we see a lot of is just systems testing during updates. What does it look like to run through an update? Let's script it out, let's walk through it. The ability to clone these live production servers into a segregated test environment, again, you can right click Clone a VM, and SAM copies aside, it's pretty instant. Whereas you start to build one up from scratch or pull from a backup, that's really a huge time savings on our end.

Rick: Now, Stan, you mentioned you test environments, which that brings up a good question. I've got to give you the nod of working in the coolest industry of all podcast guests thus far. Tell me a little bit about what you do, as much as you can and you're comfortable.

Stan: Yeah, Rick. So I work for an alternative energy company. We actually design and build wind turbines up here in Vermont. So we've got a couple different models that we produce, and then we distribute them globally. So it's kind of a unique organization both here in Vermont and globally to be involved in this.

Rick: Cool. So that probably is - I don't want to say an industry - at least a function that raises your test environment requirement compared to most small VMware environments. Is that a fair statement?

Stan: Absolutely. There are huge dev test scenarios. There are also a lot of pure development. Groups need test servers today. They needed them five minutes ago. They need a build server. They need to check out this new software. So moving towards the cloud someday.

Rick: Someday.

Stan: Someday. The cloud means a lot of things to a lot of people. We're a small organization. We don't have a huge instance of spare CPU time and disk space. But we do have a lot of users out there who need to build up these test systems, and we've got a bunch of spare servers we do that with. That's absolutely huge to be able to do that and to be able to build these environments quickly.

Rick: Sure. You say you don't have a lot of spare CPU cycles, which makes me think you might have a collection of busy applications, even though it's a relatively small environment. Is that the case, or is it a broad mix of applications?

Stan: It's a broad mix. I think our bigger issue is more of a memory management issue. It seems like everything wants globs of memory these days. We have 64, 96 gigs of memory in each blade, but still it seems like it runs out pretty quick.

Rick: Oh, you're one of the blade folk.

Stan: You know, blades are wonderful things. The reason I think we go to them is they're energy efficient. They use less power than a traditional server, especially with being able to scale up and scale down power supplies and some of the energy management features that are built into them.

Rick: That brings up a good point, because you mentioned, let's just say, a five host maximum. Do you find yourself just barely getting to the break even point on blades? Because I've always thought you needed four or five blades to break even.

Stan: That's an interesting question. I haven't done any actual math on it. They were here when I rolled in one day, and we've kind of just gone with them since.

Rick: Well, there's a strategy that people have kind of told me over the years, and that's make a decision like you're IBM or you're a billion dollar corporation or something like that. So from a compute unit, the blade server chassis is the way to go. If I had to be ultra dense, ultra big, and I needed a bunch of hosts in a rack or that type of thing, I would probably go blade. But too many times, I've not had space issues. I've not really had power issues, even though I've not really thought too frequently about power savings for blades, but that brings up a good scenario to work out. But the other one silver lining to blades is that you can have three or four blades that are ESXi hosts and then one blade that's your mail server that you intentionally don't virtualize or something like that. Do you do any of that on your blades?

Stan: One of the biggest reasons I think we go with blades is a TCO kind of concept with them. You set the chassis up one time. We have both Dell and HP chassis in-house here. You invest that time to set them up one time. You wire them up, you cable them up. Then any new blade you stick in there, it's down to a real quick setup. You image VMware onto that sucker, you bring it up. There's no cabling, there's no installation. We absolutely do run a myriad of loads on those blades. There's some RESX, some R dedicated data, utility servers for different purposes.

Doug: Have you found yourself, let's say you're obviously talking about TCO and things like that, going down looking at things that you view and stuff like that in those kind of areas, virtual desktops?

Stan: Virtual desktops are a wonderful thing. I think they capitalize on the ware we've got here with server virtualization in that you bring spare resources. You get people running Office all day long, so they're not using half of these modern CPUs. Unfortunately, in a research and development organization, you have quite the opposite. We have people who do tax their systems all day long, and they do need high-end 3D video cards to draw models or to do simulations. So I would love to have a bigger cluster, more storage, fancy toys. But VMware View I just don't think would work very well in an environment where it's so bound to the CPU.

Rick: Do you have a lot of users that are severed from the wire? Like mobile engineers or something like that?

Stan: Being in an R&D environment where you're moving around to different sites and building these turbines, we definitely have people who run around with laptops. I would guess a good 80% of our fleet is laptop at this point.

Rick: So Vermont is one of the states I haven't been to. I've been to almost every state, and especially on the east half of the country, Vermont's the only state I haven't been to. When you're talking about severed from the network users, the next best option is a mobile broadband solution. So is 3G widely available throughout Vermont? I'd almost foresee that an issue, or is it not an issue?

Stan: It absolutely is an issue. Vermont is a beautiful place, and we definitely love our environment and our open mountains. There's actually been a hesitancy to put up new towers, and as a result, I think Vermont was one of the last states that got the iPhone. It's one of the last states that's gotten 3G service. So it wasn't too long ago that if you didn't live somewhere around a city and you couldn't get cable or DSL, then you were stuck on dialup or satellite. So a working from home scenario just wasn't possible.

Rick: Because of that situation with the bandwidth, and that definitely means your users are probably best equipped with local compute.

Stan: It definitely varies. When we're going to different sites . . . I'm not sure if you've ever driven by and seen a wind turbine, but they're definitely not in downtown Boston or downtown New York where you've got good infrastructure.

Rick: Not the case, not the case. One in Columbus, Ohio just popped up on the circle highway right across the street from where I used to work. It's like in the parking lot of a car dealership. It's actually quite entertaining. The people I used to work with are a little bit entertained by it. So I know of one, let me just say that.

Doug: We're kind of similar. We've got quite a bit of a green initiative here, and there's a lot of wind turbines and things. If you look in the back of my old house, you can see one in a farm sitting on top of the hill and stuff like that. They're certainly pretty close sometimes.

Rick: What do they need to be connected? They need to somehow get on the grid, right? I guess there's monitoring solutions, so there's some sort of network, right?

Stan: If it takes more than 12 volts on the line, it's a little over my head.

Rick: There you go. If it's not AA batteries, I don't know what to do.

Stan: It's a little more involved than POE. There's a big box and a spinny meter of some kind I'm sure.

Rick: Yeah, they're heavy and expensive. That's all we know.

Doug: If it doesn't have a V in front of it, so it's not worth looking at.

Rick: So Doug, tell me a little bit about yourself and what you do there in Scotland.

Doug: Currently, I'm a systems consultant right now. So again, like Stan, I go and take in smaller enterprises, small to medium enterprises. Most of the time, looking at VMware solutions for that. So that's pretty much most of my day job is to look at the VMware side of these kind of things for these small and medium size clients. Some pretty interesting stuff. What I seem to be finding is quite a lot of these smaller business clients, they're very keen on using VMware. They're very open-minded about all the technologies that are coming out.

Again, as I mentioned earlier, they're increasingly pushed for making sure that they can test backups, auditing, and auditing those tests and things like that, and disaster recovery. So their ability to have these virtual environments, and as Stan says, the ability to possibly stand up your VMs from a backup environment is so fundamental for these small to medium enterprises. They're starting to really relish the benefits of virtualization and things it brings you, I guess.

Rick: You brought up a question to make me think a little bit. You work in that smaller environment as well, but you had mentioned VMware View in your question to Stan, at least virtual desktop. Do you find a lot of virtual desktop adoption in the small environments?

Doug: There's a lot of talk about it, and I think at the moment, a lot of the clients are still trying to grapple with it and see where they are with it. I think a lot of the clients I speak to think initially, it'll be as easy as server virtualization. But when they start to get into it a bit more and look at the storage costs and stuff like that and the kind of trade off between those things, they realize there's a bit more work involved. But certainly, they most definitely have a keen interest in it. But actually trying to put that into production, it just takes a wee big longer when everyone's got their own ideas about it and things like that. So yeah, there's some good stuff coming out of it, and a lot of people are very keen on giving it a go.

Rick: My issue back in the day with VDI was I always had trouble getting a consistent user pool that was a high number of users. It didn't make sense for 12, 15, 20 users. Even if it was 3,000. Okay, it kind of made some sense. But you almost needed a really large pool of users to make it work. I had one opportunity back in another position, but we had this hard-line application that just put a hard stop on VDI. It just wouldn't go. Tried everything we could. Probably, if we tried it again now, I was probably a little bit too ahead of the curve at the time, but I believe that solution would have been addressed by now. I think they do have something in place now with another product, but yeah, it's just there are so many factors that depend. That market has so many offerings coming up all the time. So it really, when you're shopping, makes a big difference, I think, for VDI.

Doug: I'll tell you also, as well, the thing that we have found in speaking to a number of clients is a lot of the clients, they can't quantify the OPEX. A lot of them just have usage models where every [inaudible 21:27] there's a certain amount of that that goes to an IT budget. So when IT guys are very keen in trying to get VDI in and are sure of the benefits of it, they struggle to try and put the cost together for the OPEX savings, because quite simply, the business doesn't have a model that they really analyze OPEX as maybe some of the bigger enterprises would do. So I've seen that as well. A lot of the guys are waiting to try and see what savings you can get on KPEX really before OPEX. That's something that, again, it's more like a limiting [inaudible 22:06] trying to see what savings are in the KPEX, because all our companies, they don't [inaudible 22:09] OPEX.

Rick: I think that's not limited to virtualization for desktops either. I think even server virtualization, that OPEX cost model gap is the same there as well.

Stan: Savings on the management side maybe doesn't make sense, or maybe it's not entirely clear why we're doing this.

Rick: Yeah, I don't know why I'm doing this. Why am I doing this?

Stan: Someone said it was a neat toy to play with.

Rick: Oh, yes. I heard it goes great on resumes. I'm just teasing. I love virtualization, guys. I have a lot of passion for it, as do both of you, and that's actually a good segue to the next part of the podcast. My gimmick, if you will, I'm going to ask both of you these questions. We'll start with Stan. So Stan, the part of this podcast is called "Three Views from You." I'm just going to ask you as a technologist, not necessarily related to virtualization or what you're doing now. In your past, Stan, what was the most interesting thing you've ever done with technology or story you can share with us?

Stan: So back in school, back when I was going to college, we ran a Nova shop back in the day, and I just remember being an intern with the management group at school, at college and one day coming into the office and the up down monitor saying one of the file servers was missing. I can't remember what package we were using to monitor the network, and it wasn't in a typical up down kind of language. The system's down. It was a missing kind of error message. I was phrased weird, but I think it was built into the software.

Lo and behold, the server was physically missing. A student had broken into the basement and stolen the server over the night. It was one of these big old Dells with caster wheels built into it. They rolled it away somehow. It was this great place to have a server, because it was in the basement of the administration building. The boiler room was right next door to it, so it was this great environment. 90 degrees, high humidity, and some of these servers that had been sitting down there actually had a . . . the basement would flood every now and then, so some of the servers actually had two inches of rust at the bottom of the case.

Rick: Nice.

Stan: It's this wonderful experience though. I would walk into the business and discussing where do the servers go, where do they live, what do we invest in infrastructure? Thinking back to the day when we found some space next to a boiler room.

Rick: Nice. Did they ever find the stolen server?

Stan: They did not. They had backups I guess.

Rick: There we go. How about Doug, in your past, what's the most interesting technology story you can share with us?

Doug: Before I did IT, I used to work in electronics. It basically when the mobile phone came out. If you imagine the phones that you've got now, the functionality that you have. I used to work in essentially repairing the mobile phones, and I remember getting this phone. It was probably the size of half your keyboard was actually the size of the phone. Someone said to me, "You need to fix this phone." So I'm looking at this phone, and I'm going, "Okay, fine. How do you open it?" Then my colleague bent down and brought up almost like this vise which I had to put the phone in to. He put it in the vise. This big massive thing cranks open. And it was all analog. There was no digital on this phone at all. So trying to troubleshoot that was unbelievable.

Rick: Wow, we needed a Flip cam for something like that. I'm so glad I'm equipped for this stuff now, because back then, ridiculous stuff. There's no memories. So Doug, what's going on in your current, right now, most interesting technology thing that you're most passionate about right now?

Doug: Right now, I've been doing quite a bit with Site Recovery Manager, and that's some of the stuff I really, really like, Site Recovery Manager and where it is and where it is going. There was a VMworld in Europe, and I went to that. There are a lot of nice things coming out from that. Implementing Site Recovery Manager for a few clients has been real interesting, and you have to learn a lot based on the application and how you design your VMware infrastructure in terms of being able to put VMs in the development data stores to allow a replication and things like that.

Rick: Cool. Stan, what's going on right now with you that you're most interested in technology-wise?

Stan: We're doing all kinds of things here. I would speak about the way our business works. We really leverage VMware by building up big servers and shipping them out to just kind of have utility servers on site. But it actually, I think, speaks more to the feedback we're getting. We're talking about the atmosphere here in small businesses.

These smaller groups that I occasionally work with, historically they've bought $2,500 or $3,000 servers, and what that gets you these days from a Dell or HP standpoint is just an immense amount of resources. When you put a Windows Server on that and just Windows Server, the hardware sits so idle, and people think, "This is 4-core, 6-core, 8-core, when our old server was doing this just fine with 1-core." You segue so quickly into I hear this great thing about VMware or virtualization. It's cute and adorable, but people still haven't actually embraced it internally. But then quickly after that, it becomes, how do you manage that? How do you back that up? And then, how do we really leverage that? It's interesting how quickly you move from, okay, we had this one old server that now we've overbought that now we're running VMware on. We've got this scary new backup platform.

Rick: Scary? Come on, come on.

Stan: Well, it's not scary. It's different.

Rick: I'm not cutting that out. This podcast is as-is. To all my bosses listening, I apologize. I don't screen these guests. This is the first time I've talked to the guy.

Stan: So it is scary. It's different. As the tool set changes, as you move from a host-based backup into something like Veeam's backup recovery, the tool set opens up a world of change that you can do easy replication, you can do these things that were really out of the hands of small businesses just a number of years ago. A single server, two server replication for a small business was a huge investment. Today they can get by with some low-end equipment and these amazing tools from organizations like Veeam.

Rick: Yeah. How about in the future? What are you most interested in technology-wise? What do you see coming down the pike?

Stan: It's so amazing to see what's coming out. I really hope that someone, VMware or Veeam or another organization, can bring a tool to us that allows us to maybe scale down the VDI or scale down the cloud. Some of these tools that we saw at VMworld are amazing and beautiful and wonderful, but don't really make sense for our 200, 300, 400 person company.

Rick: Yeah, I'm with you. Too many times, the small guys are left behind. How about you, Doug? What do you see in the future, most interesting, most compelling technology-wise?

Doug: I'm not sure if it's one technology, but when I at VMworld Europe, I went to a chat, and it was more about the future of cloud and stuff like that and the possible ideas. It was more thoughts out in the open rather than anything else, but the thing they described, and I've heard this a few times by various people about how each component of the virtual machine will have its own metadata, its own information about the network, the CPU, the shares on the CPU, things like that, the storage, I/O requirements, stuff like that. Basically exploding that virtual machine and putting all that information within that virtual machine so that no matter where you place that virtual machine, it will always have the same server, and it will always have the same share value network and things like that. To me, as things become more integrated versus basic around the storage site and things like that that tie into VMware, and all that metadata starts to describe the structure of your clouds and stuff like that, to me, that's really interesting, how that will develop.

Rick: Yeah, I see what you're saying, because too many times we're stuck with these vague software requirements. Gigs, megahertz, and gigabytes of disk, gigs of RAM, frequency. But how do you really quantify the needs of an application and ensure the performance? If you could kind of streamline from before the product specs to the experience and have that automated, I think you have a really good orchestration platform. You're onto something. Well, hey guys, let's wrap it up here real quick. I want to first of all thank both of you guys from being on the podcast.

Doug: No problem at all. Thanks for having us.

Stan: No problem, Rick. Sorry I had to move things around because of the weather up here in Vermont.

Rick: Oh, hey not a problem. It's been crazy here as well. We've had some weather as well. We're not quite as extreme recently. Luckily, we dodged it all. Yeah, not a problem. So everybody, if you want to follow Doug on Twitter, he is @Douglas_Carson on Twitter, and Stan is @SirStan on Twitter. Stan, Doug, hey guys, thanks for being on the show.

Doug: Thanks very much.

Stan: Thanks, Rick. Have a great day.

Rick: All right. See you guys. Cheers.

Doug: Cheers.