This week we were joined by 10-time BriForum speaker Ron Oglesby. In todays show: Ron is tearing down hardware by tearing into vendor claims about user density.
Brian: Well, hello. On Tuesday, February 28th, 2012, you’re listing to Brian and Gabe and Jack and Ron Live.
Gabe: That’s the energy that you were missing last week, man. Are you having a better day?
Brian: I had to make up for it.
Jack: We’re doing good.
Gabe: And back down to normal.
Brian: Well, I just found – for those people from San Francisco, there’s actually a little kiosk that sells Blue Bottle Coffee on my way from the subway to the office, and Blue Bottle Coffee will change your life when it comes to coffee, and it’s – I used to have to go out of my way for it, but now that I can drink Blue Bottle, it’s really, really awesome. I wonder if they would pay us, give us free coffee for that.
Ron: Is that like a brand or something out there?
Brian: Yeah, and it’s one of these things where it’s locally roasted, but they – they roast the beans. I guess, regular beans, they roast them so they have a real long shelf life, and these they don’t roast them as much. So the beans, if you buy a bag of beans, only lasts like two weeks, but it’s amazing. It’s so, so – it tastes like coffee. I feel like coffee smells amazing but tastes like shit a lot of times, and this tastes as good as it smells.
Jack: But the – but weren’t you just telling me last weekend about how the Blue Bottle people like frou-froued you because you use a drip coffee maker.
Brian: Because – yeah, I was asking about some flavor or something like that – what’s this like? And they asked, yeah, how I do it, how I make it, and I mentioned it’s automatic drip, and they definitely gave me the eye roll, like I was putting some Sam’s Club tires on their M5.
Ron: Like how dare you grind up beans and pour hot water through it, right?
Brian: Oh, man.
Gabe: And other parts of the world right now are looking at us and going this is your problems.
Brian: FML, man.
Jack: Always have to bring it down.
Brian: Well, I love that Billy Crystal from the Oscars, the one line here, which I love, is nothing buoys American spirits by – than watching an evening full of millionaires giving each other gold statues.
Gabe: Last week I tweeted something that said something like if you buy Angry Birds with Verizon – on a Verizon Android phone, do you still have to buy it if you have Sprint Android phone, and Kevin Goodman tweeted something along the lines of ah, the problems of a first world country.
Brian: So – so Ron, we’re glad that – I think is your first time joining us on this version of our show. I know you’ve been on – we’ve been doing podcasts kind of on and off for years, and I remember we did some podcasts of BriForum 2006 and you were there. Ron, of course probably most people know, but he was one of the 10-time BriForum speakers. I’m holding up your bobble head in front of the camera. Actually, we should’ve – I don't know if you’ve got a video feed, Ron. If not, we’re putting a picture of the bobble head.
Jack: Did we even mention that these are available on YouTube now?
Brian: Oh, yeah. These shows –
Jack: Not the bobble heads.
Brian: Are available on You – oh, yeah. The video shows are available on YouTube now. So this is our – if you want to watch it later and see what we look like, it’s very exciting to watch us sit around with headphones on and talk to each other.
Ron: I actually –
Gabe: Hey, Ron –
Ron: The shame is I actually think I’m gonna miss the BriForum Europe this year. I’m so swamped with everything I’ve got going on, and I didn’t have time to really come up and think of a new and exciting topic, so I didn’t want to burn any of your cycles this year. So I think this might be the first one I’ve missed since way back in Washington DC.
Gabe: You got your ten. You’re out.
Ron: I got my ten and out, right? Yeah.
Brian: Well, I will tell you that we – should you want to – I know creating a session takes a lot of effort, also, so I don’t want to sort of put that on you, but I feel confident that we can say that anyone who’s spoken at every single BriForum, you’re allowed to have your own deadline for submitting sessions. So I feel like that’s very – I feel like that’s fair. Plus, it’s BriForum, and I’m Brian. So –
Ron: Well, I’ll think about it some. This last month was like crazy for me, but I’ll give it some thought. Maybe I’ve got something cool that I’ve been doing internally that might be interesting. My latest things have been burning down – for some of our customers, has been burning down a bunch of these hardware vendors when they come out with those – oh, I got nine billion user on this server and trying to educate them.
I’m like, yeah, they ran Notepad or something, and I’ve been burning it down, especially with some of the hardware vendors because I’ve had access to that type of hardware, and I’ve been able to run Login VSI and run real tests and show that, but I don't know. Ruben and those guys and Jeroen have that pretty nailed down. So I’ll give it some thought.
Brian: Let me ask you –
Gabe: Yeah, what do you see – I was trying to put together a presentation just for something that we’re doing, and I see these – all the estimations of what – how many IOPS a Windows 7 box can do. So – and they have them – these generalized things like task worker, power user, and then I forget what the other one was, but 20 IOPS is the max that Windows 7 says it can have, and I feel like when you start generalizing things that much, you’re asking for a lot of damn trouble.
Ron: Yeah, that’s all bullshit. The –
Gabe: That’s what I –
Ron: The whole 20 – the whole 20 IOPS thing, right, what they’re doing is – and some of that’s based on some of the original work from Ruben and Jeroen, and them kind of – they have to generalize so people can wet their thumb and stick it in the air and say, “Okay, I need about this much.” So a starting point. The reality is, though, is – and it was great that you guys – that Brian and you guys put up the beverages statistical math for VDI because he hit it –
Brian: Oh, yeah.
Ron: Dead on the center, right? So a VDI instance might use – let’s say for a heavy worker – 20 IOPS. What’s the – what’s the standard deviation for that? Right? And where is –
Brian: It’s probably huge, right? At login he’s probably using 1,000 IOPS.
Ron: It’s enormous. It’s enormous. If it only had 20 IOPS, no one would ever need an SSD in their laptop, but we all know that when we slam an SSD in our laptop it rocks.
Brian: That’s amazing. There’s a tweet right there. That sums up the IOPS. Yeah. Brilliant.
Gabe: So yeah. That’s – so I mean, I guess – so is that still the baseline that people are using when they’re trying to figure this stuff out or they – are they – I guess that gets you there. Should you just take whatever you figure out for 20 IOPS per power user and then double it or triple it or quadruple it?
Ron: Yeah. I think that people are –
Gabe: Or should you just abandon it?
Ron: I think that people need a starting point to at least say what type of disk do I need. So let’s use 20 as an example for a second. If they take the number of active sessions that they’re thinking per host and they multiply it by 20 and they come up with something simple – 800 IO with some type of read/write pattern – they can at least say, “Okay, I can get X number of disk, and I’ll be kind of in that range.” Whereas if they took that same number and they talked to Cisco and some sales rep told them they were gonna get 279 sessions on the Cisco Blade that can only contain two drives, and they go, “Yeah, there’s no way I can do it there. I have to move up.”
I think the real IO problem, and it’s funny – couple of – three years ago now, when I started at Unidesk, Chris, the CTO, and I were talking about this IO issue, and for a while, we were really getting pushed by people inside the company and people outside of the company to go do a boatload of memory caching to solve the IO problem, and Chris and I, over nights and beers, just kept saying, “This is a hardware problem that we brought on ourselves by demanding cheap disk space –”
Brian: You mean Chris –
Ron: “That’s gonna be solved –” Chris Midgley from Unidesk, yeah. “That’s gonna be solved on the hardware side,” and now, two years later, you can see whether it’s kind of low-end disk like – and I hate to call it low-end disk because I actually this stuff, but less expensive disk like EqualLogic, with their XVS, where they’re dynamically moving data into SSD that needs to be there and using some of the SSD for a write cache or whether it’s the big EMC shops with fast cache and all that. Problems being solved on the hardware, you know. Throw some SSD at it to use in the appropriate places and then dump all that other stuff off on spindles, you know.
Gabe: Yeah, that’s – and there’s a lot of all-in-one products that are coming out, that are kind of niche or very specialized. IO, for instance, I think is doing that.
Gabe: So – and they’re doing it in 3U chassis that are just for – well, they market it toward desktop virtualization, but they’re for all sorts of other applications, too, but at least they’re not giant, huge boxes of iron sitting in the data center.
Ron: The real problem we’re seeing now that no one’s really talking about – I mean, I’ve had couple of tweets on it, and people have responded, but the real problem that no one’s really talking about is the enormous amounts of effing data that we’re throwing on centralized disk – which I don’t care what it is, it’s more expensive than local disk – that’s completely useless. I had a customer I was talking to on Friday, last week, that I was trying to – I was trying to basically show him ways to get around this problem, but he has a bunch of engineering students, like 1,000 of them. It’s a technical school. Their students do everything from app development to how to design a bridge and crap like that. So all of his machines require four gig, and he’s sitting there figuring out all this math, and he figured out that – it should’ve hit him earlier, but if – he figured out that every machine by default’s gonna get a .vswp file that four gig large.
Windows will default to a four gig page file, but he’s gonna reduce it to two. But if he’s got 1,000 users, he’s got over six terabytes of page files.
Brian: Of just bullshit.
Gabe: On his centralized storage.
Ron: And he told me right up. He goes, “Look. I already bought three of these XVS arrays, and now I’m having to look at buying one of these non-hybrid, just rotating, big, 14 terabyte things.” He goes, “If I would’ve known I was putting six terabytes a page file on this, I would’ve threw this shit in the trash before I ever started.” And the hardware vendors – EMC and NetApp – aren’t gonna tell you that, right?
That up to 50 percent of your space is gonna be page file hitting your disk, but I think that when you talk to some of these other vendors that have servers and are willing to look at can I leverage local disk, can we start to put this transient-type crap data on local disk to keep the cost down, that you get a little more of a responsive ear than when you’re talking to the guy at EMC.
Brian: Have – yesterday I had a couple beers with these guys from a company I’m going to mispronounce. I want to say Tegile. It’s T-E-G-I-L-E. I wanted to say t-gile. T-gile.
Gabe: I’m gonna go with Tegile.
Brian: And he says we pronounce it like with a southern drawl, and I would say –
Brian: Oh, you mean the Tegile. But he’s – he said something like Tegile. Do you know these guys, Ron?
Ron: No. No, I haven’t heard of them.
Brian: Okay, so I never heard of them, either. They make a hardware appliance for storage, and it’s kind of like what the vision is from extreme IO that they announced like a year ago and still haven’t – well, I guess they never really announced it. I kind of, to be fair, like pried it out of them, I think. But where –
Gabe: They still sponsored BriForum and didn’t tell us why – they didn’t share what the product was, but they were still out there.
Brian: Yeah, and I guess they – anyway, it’s this inline block level single instance storage. So they are primary storage, and they have NFS or fiber or whatever you want, and they can have random blocks with no hints, completely unstructured, just random blocks of data come in, and when they see duplicates, they’re gonna be able to only store one instance of the duplicate, which means that a lot of that stuff goes into memory and into SSD. Where these guys are kind of amazing – and the folks that started this company, this is their third startup. The first two they built and sold.
So they’re like – it’s like the – like Chris, I guess, also, and Don and those guys. But so they – the thing is what’s crazy about the – well, so first of all, the pricing – their box – their entry level box is $16,000.00, but that gives you ten gig – or ten terabytes of usable storage, and this is – not to mention the fact that by the time you have all this inline, sort of single instance, it might be the equivalent of huge, huge amounts, and the other thing is they have 140 of these boxes in production among different customers now. So I wonder if that is kind of this Holy Grail that I wanted because you know I’ve talked a lot about this before.
When I do – I want VDI to be persistent images because I don’t want to make people jump through all these hoops and have to contort themselves to make all the layering stuff work, though, I want to talk to you about that today, Ron. And so these – it seems super awesome, but I never even heard of these guys. I think they’ve been selling their product for six months, and I guess I’m – I don’t know if anyone else has heard of them. I guess not on the – not you guys on the phone anyway, but I guess we’ll look into that. So I guess that leads us to the next thing.
So my view of VDI is that I want VDI to be persistent so my users can continue to be local admins, so I can continue to install whatever software I want, so the users can install whatever they want, and basically that allows me to take my existing desktop, framework, structure, management, et cetera, and drop that into the data center for those scenarios where I need data center based desktops. And so I’m curious, Ron, on your thoughts on that.
Ron: No, I – I agree with the persistent model, right? Obviously I think that – I tend to think that in order to make the persistent model really viable to someone, they’ll say, “Okay, well, why would I centralize it if I’m going to patch the same way or distribute apps the same way.” It’s sometimes easy to do that. The thing that keeps the persistent model from really working – you hit on it a little bit with the storage – is that it’s storage, right? No one’s willing to pay 40, 50 gig – 30 gig, maybe even on the smaller side – per desktop on some centralized expensive storage. It’s – they’re just not willing to do it.
You’re – EMC would love that. So the trick is you’ve gotta shrink the storage down. Now, maybe it’s inline d-doop or some other model like that. I tend to still like the image-sharing model that’ll give you persistent, whatever that is. Right? Whether it’s hey, I’ve got provisioning server and I dropped RingCube on top of it, or I have Unidesk, and I’m sharing layers, but I have persistence or whatever. Because I think there’s more of a benefit there on the management side. It’s so funny. I’m not gonna say who I was talking to. I was talking to this person at Microsoft, two weeks ago, where there’re certain groups are trying to wrap their hands around what do we do with VDI. Right?
And the honest answer there was so funny. She goes, “No one here knows anything outside of their silo.” So if I’d go and I talk to the management guys, and I say, “Look, we gotta do something different for VDI. We’ve gotta –” And they go, “Well, you patch with SCCM.” And she’s like, “No, we want to do something different. Do you have any new ideas?” And they go, “Yeah, you patch with SCCM.” And she goes, “And I go over and I talk to the Hyper-V guys, and I ask them is there any way we can do this.” She goes, “Yeah, spin up a bunch of full clones.” Right? So maybe on their side it’s a storage fix. Who knows, right? But I really think that there’s two things going on out there. There’s the centralization, like you said, and the only way to make that affordable is to get the cost of storage down through inline d-doop or image sharing or whatever, but I actually think there’s a play on the management side, too.
I just think people are – I think virtualization itself, even if you do full clones like you suggest, helps with patches and stuff installing via traditional method because there’s less variation in drivers and OS builds, so you don’t have as many failures. But I really think that you take that from an enterprise where they have packaging specialists and things work in high 90 percentages to that company that’s – it’s not huge. It’s not small. It’s 550 seats. It’s 1,000 seats. And the guy who’s trying to do – run the patching is also the guy who has to monitor all the servers and make sure some of the backups were running last night, and they’re just looking for a simpler model, I think, and that’s why they’re going to VDI.
Brian: So there’s a lot to – there’s a lot to talk about in there. My – so let’s first look into the – sort of the traditional management, and I guess maybe my world has been sort of enterprise focused for too long because I’ve always sort of thought, when it comes to VDI, that if we already have ways of managing our desktops, whether it’s SCCM or what have you, that we can just sort of continue to do that within our world of VDI, and sure – and I agree with what you’re saying, Ron, that I think we’ll have more success with VDI because similar – you know, the driver bases and just the fact that all the users’ desktops are in our building.
So we don’t have to worry about sending out package and wait till Monday to find out if it ran or whatever. But maybe that is an enterprise focus thing. I don't know. Like, I don't know what the use – I don't know how many companies out there have desktop management, like an SCCM. I’m sure probably every sort of –
Gabe: Already, yeah.
Brian: Large company does but –
Ron: Yeah. Well, we run into – all the large boys have it. Even if you run into – and let’s take the extremes out, right? Take out the Goldman-Sachs of the world, where they’re talking about 30, 40, 50,000 desktops at a pop and –
Jack: Yeah, they’re like screw SCCM. That shit’s not good enough for us and –
Ron: Yeah, they’re like we wrote our own shit. Screw you guys. And then on the other side you’ve got the really small guys, right? Joe’s Plumbing Shop in Akron, Ohio. The guy who’s got –
Ron: The guy has 20 desktops total, and a couple of his guys have laptops for mobile quotes or whatever. But take those – take the in betweens and even when you think about government – government’s really interesting, and Brian, I know you did a lot of government work when you were at HP back in the day. But we think of government as these millions of desktops, but they’re all managed by these little groups. You go into the military or DODD or you go into the US Census Bureau or you go into any of these things, and they all talk about I manage my 500, I manage my 1,500 desktops, I manage my 4,000 desktops, and that’s where they feel that pain. Right?
It’s all these redundant – they have their own little fiefdom; they can do what they want; they have to live up to certain codes or regs or laws that say what they have to do security-wise or reporting-wise. But that poor guy, he’s got 1,500 desktops to deal with, and then he’s got all these other things, too, and quite frankly – you should know, Brian. You were an SMS geek way back in the day. How hard is it to package and become a really good application packager or really good at distributing patches. Right? It’s not an overnight thing in most cases.
Brian: Now, on the flipside of that, though, sort of by definition, whatever they’re doing now is working because their company’s in business. They have users.
Ron: Oh, yeah.
Brian: So I’m nervous. One of my big fears about moving to, I guess I could say desktop virtualization, but specifically VID, is that VDI is such a change, just in terms of the general user experience because you have – you take the desktop away from the user and centralize and then have to deal the different devices and protocols and that kind of stuff. So that’s a big change right there, and then at the same time, if we also sort of fundamentally change the way that we’re managing desktops, that’s a lot to change at once.
And my fear is a lot of times VDI projects either get stuck in pilot or fail outright, not because the VDI aspect of it failed, but because this massive changing trying to go whether it’s a shared desktop model or something like that. That’s the reason that VDI fails.
Gabe: Right. So rather than changing it all at once, you can – if you can get your desktops into the data center and deal with that first, while still managing it the same way you’re currently managing your desktops, then that at least lets you take a phased approach to utter completion – utter VDI bliss, if such thing exists.
Ron: Well, and I – see I think you hit on it a little bit, but I would flip it. I would say that there’s a lot of people that make the decision to go to VDI – kind of like you were saying, Brian, that there’s all these things going on that – which protocol do I use, and how do I get them remote access, and what happens if the network link is down to this remote site and all that stuff, and I think they go there almost in spite of that. Right? They’re looking, a lot of times, for a new way to patch or manage. I had – got a customer. He’s a smaller customer in San Antonio, about 500, 800 seats, something like that, but he’s like 15 locations, right?
And he has all these locations with 20 meg – 20 users, 50 users, 80 users, and for him it was really I’m sick of driving out to site whatever every three days to fix this guy’s computer and to do that. So it was really he was looking for a new way to manage. Now the protocols and thin clients and all the other stuff he had to deal with, I think if he could’ve managed like he did in VDI, some centralized way reset desktops, push application, I patch some image and it – I don't have to know how to package. I just patch and it goes out, and he didn’t have to VDI it, I think he would’ve done that.
But there’s really no good tools that change the way – there’s nothing really that fundamentally has changed the way that we patch Windows, that we update applications, or install applications, and I think there are some users that are really looking for that.
Brian: Now, in that use case, for the San Antonio customer, why not something like a MokaFive or virtual computer or Wanova or something like that?
Ron: Um-hum. Well, there he would – he would’ve – and those might have solved his problems. I know that he would’ve had to bought new hardware and went with a new model. Right? Whether it was a type 1 hypervisor or whatever. He was kind of sold on the idea of bringing as much as he could centrally, buying some servers, spending the money there, and then he – they do have thin clients, but they also have some repurposed desktops as thin clients, and he went that direction. Maybe a distributed model would’ve done that. In our case – in his case, specifically, most of his desktops aren’t even persistent.
I would have to say 70, 80 percent of them aren’t even persistent. It’s – vast majority of it, for him, was applications updates, desktop repairs, and desktop patching. So maybe a distributed model – if type 1 hypervisors would’ve got where we thought they would all be two or three years ago –
Jack: I wonder if there’s – if you look at a distributed solution that would maybe require that you refresh your hardware to be able to – your end user hardware in order to support type 2, I wonder what the – taking into account everything that we say about cost models, I wonder what the cost model of that endpoint refresh looks like versus going with a centralized solution.
Ron: Yeah, and I don't know, right? I – we have to know that – you’d have to have decent hardware. It probably wouldn’t have to be rock star gaming system type hardware to do it. You’d have to have decent hardware to run a type 1, and we never really know because there haven’t been a lot of financial models built around that saying, hey, what’s better for me: VDI or the type 1 distributed with management pieces? Really because those type 1s all burn down, when compared to the expectations everyone had for them. They really burned down before anyone got that far. Windows 8 –
Brian: Those with VDI?
Ron: Who knows? That might change that.
Brian: Doesn’t VDI burn down before people get too far along to –
Ron: No, VDI just kept kind of smoldering, I would say. Right? Like every – it was like the guy who doesn’t know how to start a fire. He throws three logs in a pit, and he tosses some gas on it and a match, right? And it flames really big, and then it kind of dies down, so he throws more gas on it, right? He just keeps doing that over and over until the logs catch.
Brian: So –
Gabe: I watched a guy at a campground do that without the gas.
Brian: Regardless, you end up with a pile of ash, just one of them slowly smoldered while the other one flash fired and then – so I wonder – and there’s maybe something to be said here, also, that if you’re going to have to spend the money maybe there’s a perception that doing something that’s client based is the sort of “the old” way of doing things. I’m not advocating this, by the way. I’m just kind of going through this thought exercise where you say, “Well, if I – is my use case supports allowing VDI, then I also have the flexibility of users to work from other client devices and pop into any computer on the planet and get their desktop and the central backup and that kind of stuff.”
So maybe there’s something like, okay, so there are good future ways to do distributed managed desktops, and that’s a type 1, type 2 hypervisor on the client, something like that, and there’s the VDI way. So maybe there’s something to that. Like, well, if I’m going to reengineer what I’m doing now, maybe I should just go with everything in the data center and kind of say goodbye to the managing these endpoints out there.
Ron: Yeah. Yeah. Endpoint management’s always been a pain in the butt, all the way back to the late 90s when we all thought ASPs would be the end of the world. But what I see is a lot of people – and it’s kind of – it’s funny because I remember seeing one of your tech target spiels a couple years ago in New York, and one of the things you were harping on, and you probably still do it – it’s probably the same spiel – but one of the things you were harping on was –
Ron: VDI – VDI doesn’t cost the same as the $300.00 Dell you shove under your desk, just like you used to be able to get e-mail in some crappy interface in Windows 3 or Windows 95, but now you get it on a Blackberry and an iPad, and it does all these other things, so it costs more. That’s why we went out and spent a bunch of money on BES servers back in the day, but I see a lot people – when I go in and I talk to them about what are you trying to do, they say that. They say, “Look, we’re gonna build VDI to stop managing these desktops this way, but the cool thing is we’re going to start testing remote access and bring your own device and all these other things.”
So they’re really starting – they think VDI is that stepping stone to allowing them to do that.
Brian: And it’s – it’s interesting, though, because they’re not – fundamentally VDI does not remove their need for having to manage instances of Windows. I think that that’s one of the big misconceptions. Maybe they can – maybe Windows becomes easier to manage because, as you said earlier, if they’re all VMs they all have the same sort of cord driver set, and they’re all in their own basement.
So they’re easy to update and stuff, and I guess fundamentally – my feeling was always if your VDI environment is able to work off these shared non-persistent models, then maybe you can get some management savings from that, but you still have to deal with applications that are not compatible and user environment and all that kind of stuff. So there’s no way you’re leaving the world of Windows management at all.
Ron: No, there’s still a lot of Windows in there. That’s for sure. I – you would be surprised at the number of calls that I get or the support guys get on basic Windows things. Just – we’re Unidesk, right? We’re there for their apps and building a desktop, and someone will call up and say this guy can’t do this, and we’ll wind up Googling the error for some Microsoft knowledge base article on they didn’t install something in their Windows package. So they still wind up doing that. I think that VDI without a management system – think about it. If you had View or XenDesktop with no image sharing, would anybody move to it?
Ron: Well, yeah. But they did their whole weird thing with individual terminal servers and whatnot, right? But yeah, how many people would be flocking to it if we said, “Well, here’s what you’re gonna do. You’re gonna make these things VMs. You’re gonna run SCCM on them, and you’re gonna have to have a guy that can package apps over here, and you’re gonna have to have this guy that can –” Some might do it for remote access, kind of along the lines of last week where Claudio was talking about these people wanted to remote access their desktops but not a lot, right?
A lot of people are – they’re really looking for a different way to update desktops. They’re still gonna have to update them, but they want to make it simple.
Brian: Now –
Gabe: Well, sure if it was that alluring to do persistent stuff or static images without any other built in management, everybody would be doing it, but it’s too complex and not worth the effort.
Ron: Yeah, people – yeah, and it was funny. I always talk about View Composer because I was – I was at GlassHouse and then Dell – big VMware shops, and I was at a meeting at VMware once where they introduced the concept of Composer before it became an actual product, and the straight out – the solution was to do two things and that – we want to reduce the storage footprint, and we want to make it so you can patch an image once and push it to everybody. We want to make it so that you don’t have to package a fix, you don’t have to use w-sauce, you don’t have to do any of these things.
You patch this image once, you say go, and it goes out, and it does those two things really well, right? We’ve seen – they had to go out and get a product, and they had to buy RTO, and they had to do all those things, but that’s – that’s what they had to do to make people buy VDI because if you just moved a bunch of thick images centrally, people didn’t – the admin guy didn’t see that it made his job any easier from a management perspective.
Brian: But now, Composer, to me, seems like the most craziest – well, amongst the crazy products in the world, Composer is there, and this is not – it’s the same with clones, and a lot of this kind of stuff – provisioning services – there’s – and actually the new hypercache stuff from Quest Software, which only work – all of these products that require a shared master image and then they make all these little – differential files for all the users, so they can have the personalization that each VM needs to physically boot up and have a different identity and stuff.
But when you – when you update your master and then recompose, you blow away all the settings changes for all those individual users, and so that just goes back to my – one of my arguments I’ve been giving for a few years. So Composer only works for non-persistent desktops. So if you’re doing all that, I mean, geez. Why aren’t you just using Terminal Server and doing the whole thing with 3X of user density?
Ron: Which – yeah. Which is what I say a lot, right, is that it’s great for a non-persistent thing, non-persistent workspace for a user, but because it was built on the block level, everything – that – think about it. Because of the way Composer was designed, or any block-based image sharing was designed, you have to go out and get an application tool, a profile tool, AppSense, RES, RTO, whatever it is, or else it’s a terminal server with roaming profiles.
Brian: And way more expensive.
Ron: And way more expensive, right? Your density is down and –
Brian: And stupid ass licensing.
Ron: Yeah, and crazy Microsoft licensing.
Brian: For Microsoft. Yeah. Yeah.
Ron: And the funny thing is I always start off my presentations, when I do demos, explaining that to people, when I’m just talking about Unidesk because that’s one of the things – we didn’t actually want to be in the non-persistent business. I was talking to someone yesterday at Unidesk, bringing them up to speed. When we first rolled out the product in beta and version one and whatnot, there was no non-persistent desktop in Unidesk. We’re like, well, that’s covered. Go buy Terminal Server. Go buy View. It was later, after people bought our system and used it, that they said we want to do this with our non-persistent desktops. Now, while I may think that if a desktop is non-persistent you should first try it on terminal services and then use – if a customer tells me he wants to buy a product for non-persistent, we’re gonna make it for him, but you’re dead on. If you can do non-persistent try RDS first, then move to a Composer or something like that, and what we found was that they weren’t coming to us because our image sharing was necessarily any better. It was that we didn’t do it at the block level.
So all those things that you talked about, even our non-persistent desktops, the computer name, the domain SID, all that stuff that happens when a machine is first built are retained. We don’t try – we try not to kill all that stuff that makes a desktop kind of what a desktop is.
Brian: Oh, and so I think you answer my – a question from earlier because you mentioned this customer in San Antonio with 70 or 80 percent of that person’s VDI desktops being non-persistent, and I was sort of gonna ask you what the hell.
Ron: Yeah. Yeah. Well, it was about – it was kind of two-fold. It was about the applications, right? He didn’t – he tried some ThinApp. They looked at other things. It never – he never found anything that just worked for him, and of course, for the 20 percent that we’re persistent they needed to be persistent, but the difference with us is that since we’re not doing it at the block level, since the OS and the apps and everything is in a layer, essentially, you can retain all that stuff that you were talking about.
You can retain the computer name and the domain SID and freaking when Symantic AntiVirus boots the first time and it writes a UUID into the registry and registers with the Enterprise management server, it all looks and acts like it should versus a recompose, where essentially the desktops are blown away and re-created with the same name.
Brian: So let’s talk a little bit – I know that – so you work at Unidesk, and I know you’re presence here on the show is sort of not related to Unidesk but more of like Ron the guy who’s been in this industry for – I don't know 10 plus years, 15 whatever it is – and doing Terminal Server for 10 years. So I know your background is consulting and been Terminal Server, kind of VMware, guy. You and I wrote a book about Terminal Server together in 2004. You’ve since written – what – two, three, four books on VMware. You spent a short time at Dell, but you’ve been at – or I mean, you spent a year at Dell. The first jump on the sort vendor/provider side. Now you’ve been at Unidesk for coming up on two years now, right?
Ron: Yep. Three.
Ron: It’ll be three in October.
Brian: Holy moly.
Gabe: Is this the longest you’ve been with a company?
Ron: No, I was at RapidApp for seven years –
Ron: Before we got bought by GlassHouse, yeah. So I’m hoping –
Brian: That’s right.
Ron: To get seven here. We’ll see.
Brian: So but kind of fill us in because I don't know if we’ve actually talked about Unidesk on brianmadden.com in the past year or so, and so maybe just kind of give a primer for those listening who are not specifically familiar with Unidesk, but then I want to sort of dig into what’s changed since we last talked, and that could’ve been like a year ago, even. I don't know.
Ron: Yeah, so we’re a layering company, and layering, to us and to most people that are really into VDI – that pay attention to it – is really kind of the next – I would say a guy a VMware told me it was the next gen of what Composer should be like, and instead of doing provisioning and applications and stuff with all these kind of point tools, we start at the OS, and instead of doing, hey, I’ve got this VMDK and I’m gonna share it, so I’ve got all these delta disk and it’s block-level sharing, we do it at the file system and registry level. Essentially layering, whether it’s RingCube or us or whoever, is file system and registry virtualization, the ability to merge those and present them to Windows so it sees a C drive and it sees a registry, and it just looks normal, but underneath you have an OS layer that contains all of its file system and objects and registry entries. You have application layers that might have been built on top of the OS layer, but they’re unique. They only contain their file system objects and registry entries, and then each desktop has a personalization layer that contains the things that get changed for that desktop, whether it’s a profile change, whether it’s a programmatic change from antivirus, right, pushing out a UUID to the machine, or whether it’s a user-installed app, and what we do is we replace those tools.
Like in a VMware environment, we replace – we would replace Composer. In most cases, we would replace ThinApp, and of course, you wouldn’t need roaming profiles or RTO or anything because we would provide persistent desktops that are sharing images and can be single image updated and all that crap, right?
Brian: So the idea, or the goal, of layering is the sort of user experience of persistent desktops, where a user makes a change, installs an app, what have you, and it sticks until next time they’re there, but with the sort of management simplicity and storage efficiencies of the shared non-persistent desktop.
Ron: Yeah, yeah. It’s kind of a merging of stateless and stateful environments. I didn’t –
Gabe: That’s what we always wanted shared-image VDI to be. The – is –
Ron: Yeah, exactly. We get a lot of people to ask us about that. Now, we’re gonna try to take it to the next step, and we get a lot of people asking us to do some other things, but I always tell people – and I used to tell this during my demo – but I didn’t come to Unidesk because of VDI, and when they first reached out to me, our chief marketing officer, Tom Rose, reached out to me a couple times. I actually blew him off. I was like, ah, I’m not all that into VDI, whatever dude. It’s centralized desktops.
But it was when I talked to Chris, the CTO, and we really dug into the technology, and one slide he showed me the concept of, okay, if I separate these things into logical items that we can then manage independently, and I do this stuff with the user personalization layer, where I can actually control some of the data and where it lands – the cool thing about is I can do repairs, and he showed me the concept of what if I had a central Window 7 or a central set of applications that I could push to VDI and I could push that same image down to a desktop and it ran there? This, of course, was when we all thought type 1 hypervisors were right around the corner. Now I’ve got an IT platform where I can manage desktops, laptops –
Gabe: Did we lose – did we lose him?
Ron: The central VDI images –
Gabe: There he is.
Ron: The application packages are all the same. Because we can do drivers and all that stuff, and by the way, if someone toasts a laptop, you hit rollback, and it takes them back to yesterday in their personalization layer. So that’s – that’s kind of why I went here was because of the layering technology, not necessarily because of VDI.
Gabe: So is that still in the cards, as far as the desktop management, even on the physical side – even – in light of the fact that type 1 hypervisors never took off? Is there anything –?
Ron: Yeah, it’s still something we want to do and we talk about all the time. Generally, the discussions are around what platform does it land on because we’re not in the business of making our own type 1 or anything like that. Will Windows 8 change that with its own built-in hyper V? Maybe. Is it the right platform? Will the SDK and everything be there for us to automate it and do what we need to do? Who knows, right? It’s not even there yet.
Brian: Because, fundamentally, the Unidesk layering, that’s happening at the hypervisor level. So it’s the hypervisor doing all its magic of melding all these sort of different disk images together.
Ron: Kind of. It’s above that. So all the VM sees is disks attached. The actually merging of the disk is happening from within the drivers. We call it the CFS, the composite file system, that actually merges it. So a little trick, you could actually go into a Unidesk desk and list the disk volumes, and you can see multiple disk volumes, but if you just go to Windows Explorer and double click on the C drive, all you see is a C drive. So it’s actually kind of happening below the executing environment at the driver level but above the hypervisor. So we can actually run our stuff on almost anything, hypervisor-wise. We just have to write to their SDK to do copy files and stuff like that.
Brian: There’s a bunch of questions that are coming in, and we’ll get through these, but I’ll interject one that – so you’re mentioning hypervisor – like, fundamentally, the hypervisor’s technology agnostic, but one of the users is asking about Hyper-V support from the Unidesk product.
Ron: I don’t have a time for that. It’s actually the next major thing on the list. We’re making an architectural change here, right now, that’s gonna really even allow for that to run, but we get asked about that all the time. We used to get asked about XenServer a lot, and that has completely fallen off. We are actually getting pushed from customers asking about Hyper-V and from a couple of our bigger partners really wanting us to do it with Microsoft. So that’ll probably happen. I don’t have a timeline. It’s not gonna pop out in July or anything, but it definitely is gonna happen.
Gabe: Is that a – it would make sense to probably target that for a Hyper-V V3, to work with Hyper-V V3.
Ron: Or something, yeah.
Gabe: Yeah. Just – there’s so much buzz around that, and there’s so much excitement that Microsoft is gonna get it right this time, or more right. I guess some people are using it now but – yeah.
Ron: Well, and is it not a ripe market where VMware has Composer, Citrix has Provisioning Server, MicroV – Hyper-V has essentially – they’ve got some stuff going on with Quest, but it’s not got a great adoption. So there’s – there’s a more of an open market there than there is even is with VMware or Citrix.
Brian: That’s a good point because – yeah. So okay – so this layering stuff – and we’ll get through the – some of these other questions that people are posting, and feel free to pop more into the chat room if you’re listening. Let’s take a step back though because the first – my first exposure to this concept of layering was from Chetan at Atlantis Computer.
Brian: Let’s call it three years, four years ago, and Atlantis was talking early on about this layering and how all these layers – different layers – could be patched to different levels and then re, sort of, combined into desktops and stuff, and there’s a lot of logistical problems with that, and Atlantis since has gone on to not focus on layering but focus on performance, and Atlantis is doing like – they’re doing fine as a company, but they sort of didn’t go after that layering, and I think part of that was because this layering is just complex because if you’re changing things within layers –
Like maybe you swap out something at the desktop layer – level – and it now has files that conflict with a higher level layer or user installs something, and it writes into their layer, but then the desktop – those components that need to be in the desktop layer that the user couldn’t write or didn’t persist. So my general fear about layering was that it was an amazing, awesome concept, and I want like crazy for it to work, but I’m just scared that it’s too brittle and that the exceptions would add up, and at some point I’m just like, well, eff this. If I would’ve just gone to persistent desktops in the first place and then had like a CCM or an Altiris, I wouldn’t have to deal with any of this shit in the first place.
Ron: Yep. I – I used to worry about that a lot, too, and actually it’s funny. The guys that wrote in are – one of our kernel engineers and our chief architect and Chris, our CTO – those guys noodled for a long time on the concept of – because we stack layers. We’re not just like RingCube that goes on top of a block image. We’ve got layers on top of layers, all the way up to the personalization layer, and they noodled for a long time on all the conflict that could happen, right? The – and how do we resolve it? And they built stuff internally for layer priorities.
We still – when you’re doing – when you’re creating layers, we still respect SMI rules and all that stuff. So all the stuff people are kind of used to using still works. The interesting thing was they even built in some stuff that administrators can use to change layer priorities. So what am I gonna do with app has DLL 1 and this has DLL 2 and this one needs to overwrite that one? So we even put in stuff that users could flip flop layer priorities, if they needed to, in the field. The funny thing is – like I said, I’ve been here two and a half, going on three years. I’ve used it twice in all my time here, and I’ve probably talked to every customer, demo or production, that Unidesk has, and I’ve got customers that – 1,500, 2,000, 4,000 seats – that have never used it. So it’s like the under –
Gabe: So when you say the users – I’m sorry. When you say the users can flip flop it in the field, do you mean like the administrators or the users.
Ron: The administrator. The administrator. I’m sorry. Yeah, the administrator.
Gabe: I just want to make sure.
Rob: So the conflict solution is really – you’re dead on Brian. It’s a big worry, and it’s actually really important. You have to have the mechanisms underneath and actually be refining them, right? Because day one they’re not gonna be perfect, but be refining them as you go to make sure that the desktop operates kind of as expected, and then we also put in another technology, as someone in the list asked a little bit about RingCube, and I think this is where there’s gonna be some issues until Citrix works out a bridge between the block-level provisioning and the RingCube layer. We also put in a technology that administrators can override things in the personalization layer.
So when someone does call up and say, “Hey, Office stopped working.” That’s all they know. They don’t know that it was the stupid plug-in they downloaded or whatever. When an administrator checks a box, he looks at that desktop and checks a box and says, okay, override anything that they’ve done that’s conflicted with the office layer, so DLLs, registry entries, they’ve overwritten things, whatever, and it essentially overrides that way. We don’t do it by default because, hey, they might’ve installed a plug-in that replaces a DLL, and it works just fine.
So the idea is to be able to repair it when it happens really quick, but the key is the underlying system that allow these layers to merge and, like you said, not step all over each other.
Gabe: Do those repairs happen on the fly, or does that – are there reboots in between?
Ron: It depends. Typically –
Gabe: Where the –
Ron: It’s a reboot in between. Yeah, because if they installed something and it overrode C program files common blah, blah, blah, dot DLL, we’re just gonna – and it’s a restart is exactly what it is. We say – we call it punch through, where the app is basically punching through anything that the user has done. We eradicate their changes to those DLLs or registry entries, restart the machine, and it’s back up to where it was. Now, nothing else they did was lost. The WinRAR they installed, their profile, the iTunes they have, all that stuff’s still there but just anything directly conflicting with that specific app layer.
Brian: So would you characterize my fear of layering as sort of not really –
Brian: Yeah, maybe unfounded. Would you –? Yeah.
Ron: You know what I would – you know what I would liken it to? I would liken it to the early days of server virtualization, where everybody was worried: what if one server ate up all the resources on the host? Right? That was a big sweat. That’s why no one wanted to move out of test. That’s why you didn’t move it out of the labs because – ah, you don’t understand. My server really – is there a potential for issues? Sure there is, right? But the underlying hypervisor in that case, they built the scheduling mechanisms and the throttling mechanism, so that if it was only given two VCPUs it could only eat two cores, and there were 20 other cores in the box, et cetera, et cetera.
It’s a valid thing to bring forward. If the technology has a way to handle it, then great, right? I don't think all the technologies have a great way to handle it. I think we have a pretty good way of handling it, and we continue to refine it, but it’s a valid concern that we feel that we’ve addressed.
Brian: So there’s a bunch of questions, and I don't know if you can – I don’t even know how to ask that. So I see Dan Brinkman and some anonymous people are just asking about sort of Unidesk versus RingCube or AppSense or Winova and these other kinds of things. I don’t know if you can even share any thoughts here. I don’t want to get into why one is –
Ron: Who’s better? Who’s thumping chest. Yeah.
Brian: Yeah, but what – I don't know.
Ron: Well –
Brian: What does Unidesk do that’s really cool?
Ron: Yeah, well, when we first came out, and you guys even mentioned it before, I pushed on user-installed apps when we first came out because no one else was doing it, which we can support the user-installed apps, but the reason I pushed it was because we wanted a splash, and then you guys mentioned, hey, now you guys are doing – you’re getting away from persona. You’re more doing apps and stuff, and I think that’s – the really cool thing is that we handle the user-installed apps and anything the user wants to do, but in the same interface and the same technology – the user doesn’t have to go to anyplace else – he can package or layer his apps, and they don’t have to learn anything, right?
They don’t have to learn about isolation or bubbles or do I have to link these two packages, these two bubbles, so they communicate, or I can’t install Adobe, the full package, because I can’t support drivers, right? So that’s really been a big thing, as of late has really been the application stuff.
Brian: And you mentioned earlier that, fundamentally, there’s nothing really preventing your stuff from working in a physical desktop environment, which I could extrapolate out to say if you can make it work in a physical environment, then it could work in sort of any virtual environment, client, hypervisor, streamed, OS, any server platform. So if a company wanted to use a Unidesk for sort of all their Windows desktops in their environment, regardless of where those are, that’s not there today but sort of – would you call that on the roadmap, or would you call that on like the concept of what could happen in the future?
Ron: That’s the long-term goal today. When we talk about it, we always talk about we’re in VDI today. You’ve gotta make your money somewhere, and that’s where the change happens to be, but the long-term goal would be to do this for desktops, too, right? One management platform or interface for the end user that can do physical desktops, also. We get asked a lot about servers. Can you do server OSs? There’s nothing technically that would keep us from doing server OSs. We just haven’t put the focus on it and put the engineering time in on it, and we’re probably not going to.
Brian: Imagine user-installed apps for servers. The user in this case is like Joe junior admin. Okay, let’s change topics here because we’ve got about seven minutes left, and I want to mention – a couple things other I want to talk about today, and I guess I’ll first mention it as continuation. A lot of the layering conversation we’ve had today, and the challenges with clones and everything, is it’s a topic that we cover at great length in our upcoming book.
So Gabe, Jack, and I sort of put our heads together and are writing the book on desktop virtualization that, frankly, we’ve wanted to write for probably ten years, and so we – I’ve written books before and Ron, of course, books, but all of our old books – and I think, Ron, this is the case for your old books, also, and certainly the ones we did together. We were very sort of product focused. We never wrote instruction manuals. We sort of wrote like if you’re the VMware person at your company or you’re the Citrix person, here’s what you gotta know in order to roll out Citrix or whatever, and those books were huge.
They were like 60 or 70 bucks to buy, and they were 900 pages, and they were insane. So what we writing right now is a book sort of on the whole concept of desktop virtualization and why it didn’t take off like everyone thought it would and what you can really do about it and how to think about VDI versus client hypervisors and applications as a service and Windows’ future apps and web apps and all that sort of stuff, and this is the book that we’re – I’m actually gonna go home after this, and hopefully, we’ll finish writing this thing today, but I think it’ll be available on Amazon in the Kindle form in the next couple weeks.
Brian: It’ll be like 200 pages, instead of 900 pages, so I think we’ll – I think we’ll be able to hit like the $10.00 price point on Kindle probably. We do it like as a hardcover, sell it for 25 bucks in paper form. Our working title – I was asking on Twitter earlier if anyone saw. Our working title is either called The VDI Delusion or possibly The VDI Fallacy.
Ron: You guys are haters. You’re VDI and Microsoft haters.
Brian: Well, and so – if you look at – so the – like The VDI Delusion is cool, but Richard Dawkins wrote that sort of pro-atheism book called The God Delusion, and I don’t want people to sort of be mad at me because they’re – don’t – like I don’t want to like bring any religion correlation into this and – but if you look at the definition of fallacy, it’s – the definition literally says, “it’s a mistaken belief, especially one based on unsound arguments.” So I feel like –
Gabe: So when I heard – or when I heard delusion, I did not think for one second that it had anything to do with religion or The God Fallacy or anything – or Delusion or anything like that. Now, fallacy, on the other hand –
Brian: Different word.
Gabe: The first syllable of that –
Brian: Different word. Different spelling.
Gabe: It’s the first syllable that gets me to hung up on that one.
Brian: So delusion is an idiosyncratic belief or impression that is firmly maintained despite being contradicted by what is generally accepted as reality or rational argument.
Gabe: So if we put that definition on the first page, too, I think we’re good.
Brian: But in the – I think the subtitle will be something like Why Desktop Virtualization Failed to Live Up to the Hype and What the Future Enterprise Desktop Will Really Look Like. So this is great because we do get into why – and oh, what did I – Risk On/Risk Off tweeted the title should be called The VDI Delusion: How One Company Flimflammed an Entire Installed Base Into Buying Shelfware or something like – something like that, and we do open up, and it’s pretty critical of VMware early on, that how they were successful with server virtualization and why, and then how they tried to equate that to desktop virtualization and how Citrix got involved, then like the reality VDI where it didn’t live up to the hype, and we get into our cost analysis and how to lie with cost models, and we do talk about where VDI works and where desktop virtualization works because it’s certainly – I’m not truly – I’m a big hater of VDI when it tried to be pushed into scenarios where it doesn’t make sense, which means that people hate VDI when really it’s like they should hate the person who recommended it because it didn’t work, and I point out we do have, what, 5 or 10 million VDI users in the world who are all loving it.
So there’s certainly a use case. It’s just not – we’re not turning all of our users into VDI users, and then we get into the future of Windows applications and WinApp stores and all these alternative apps, device native. I think it’s a fun book, and we kind of split the writing duties. So I wrote a bunch of chapters. Jack wrote some chapters. Gabe wrote some chapters. We’re sort of stitching it all together now, but The VDI Delusion, The VDI Fallacy – look for it. It’ll be Kindle hopefully in the next couple weeks –
Brian: And books – on Amazon, I guess, probably in another – end of the month.
Ron: You guys are bringing Jack right along, right?
Ron: Bringing him along, huh?
Brian: Yeah, and for this – for this book, it’s great, too, because there’s a lot that he’s able to contribute, even really only focusing on the space for – it’s been six months, I guess, but Jack wrote the entire chapter that’s about these other non-Windows application platforms. So how you deal with all these native apps in – well, I don't know – IOS and web apps and all –
Jack: All the things that these punk kids do.
Gabe: Like Jack.
Brian: And –
Gabe: You know, it’s really hard. The challenge with Jack is giving him 15 years of accrued cynicism in such a short period of time, and I think he’s taken to it well with the Microsoft stuff, at least.
Brian: Well, and now – and especially – and I think I mentioned this before, but we brought in Jack around the time that we sort of formally launched our consumerization efforts, and I would argue now that Jack knows more about MDM and mobile app management than I do, and he’s actually doing expert videocasts for us now, and people are asking for him, not me, because Jack’s writing – he’s writing three articles a week on consumerzied IT, and I’m writing like once every two weeks on that side or whatever. So yeah, it’s working out well so far. So that – I wanted to get that out there and a lot of stuff we talk about is in that book.
So I’m really – this is the book I’m actually most proud of. I should maybe take that back. I don’t want to say that, but this is – this is a book that – it’s cool because it has like my most knowledge of the industry in it, as opposed to – it’s not like a product guide. It’s not product specific. Probably the word XenApp and View are only mentioned half a dozen times in the entire book. It’s more about sort of the how you deal with this technology and how you think about delivering Windows desktops in your environment.
Jack: I see we had a suggestion from Dan Brinkman for VDI#fail, which is – which was the title idea that I was going for but –
Brian: Yeah, I – yeah, Jack wants it to be called VDI –
Brian: Yeah, that’s what I was kind of thinking because to me, like, if you call it The VDI Delusion or The VDI Fallacy, that makes a statement, and VDI Fail makes it sound too trivial, like the punk kids, FML kind of first world problem kind of thing. I don't know. So we’ll kind of see where we end up with that. The other thing I wanted to mention is – so we’ve been talking on Twitter this morning about Noah Wasmer leaving VMware, and so Noah – we’ve mentioned him before. He’s been on podcasts before and videos we’ve done. Noah was an old-school terminal services person from ten years ago. He joined VMware, went into their View group, and then he’s the guy who sort of created this concept of VMware Horizon, which is their platform that – and product and service that will tie together web apps and web app provisioning and consumer apps and Windows apps and everything.
And VMware’s going big. I feel like VMware’s almost like dropping View from their sexy list and replacing with Horizon, and Citrix is doing this, too, with what they call their Cloud Gateway, and I think, what, Centrix does this, and there’s a – Quest is going down this path. So – but he’s Mr. Horizon for VMware, but it was announced – by the way, actually, the way I found out he left VMware – and I’ll give a shout out to this website. It’s called – wait – Nimbil? Shit, I just forgot right now. Let me look. N-I-M-B-I-L. So it’s this – yeah, it’s called like the social CRM platform. I just used a free version.
You give it access to your LinkedIn account and Twitter, and like it can login to your e-mail if you want to, and it mines all that for data, and for me it sends it – like the best feature on it and this is – what I used is free – is it sends you an e-mail each day with all your LinkedIn contacts who changed jobs, every day. So it’s like please congratulate the following five people, and like today, it’s like Noah Wasmer, VP Products at MobileIron, and I’m like, oh, holy shit. So, A.) we know it’s public because it’s on his LinkedIn profile, and so I tweeted about that, and a bunch of people have been sort of retweeting and saying it’s a loss for VMware and everything. So he’s taken this week off.
He’s a recent and proud father, and he – so he’s taken this week, and he starts at MobileIron officially next week, but I said, hey, man, give me a sound bite for our show, like as to why you’re leaving VMware, and as a matter of fact, I actually just got this e-mail during the show, so I haven’t even read it yet. So here’s what he says: “I loved my five years at VMware –” I have to magnify this. I can’t see it. “I loved my five years at VMware –”
Jack: Make sure at the end of the e-mail it doesn’t say, oh, by the way, don’t mention this on the air.
Gabe: Yeah, pre-read.
Brian: “I loved my five years at VMware. I think VMware is going to drive some amazing innovation this year around a new way of working. For me, the fastest, most aggressive, and exciting changes are in mobile IT, IOS, Android, and even Windows on ARM (we can debate the enterprise use case). MobileIron is blazing a new ground in the space, and I’m thrilled to be coming onboard. Okay, well, he’s – that’s a good VP-level statement of non-information. Love VMware, this is exciting, blah, blah. So he’ll be back from vacation next week, so maybe we can try and pull him in.
Gabe: And I guess we have another MDM company to add to the list, right?
Brian: Yep. They’re on the list already, aren’t they? Yeah.
Gabe: Are they?
Brian: MobileIron, I think, is turning it up. They’re a – people say good things about them. So alright, I think that’s about – we are like a little bit over our hour now. I don't know if anyone else has any final thoughts or anything for today.
Gabe: No, not really. Glad Ron could join us.
Ron: Yeah, this has been a lot of fun.
Brian: Yeah, well, Ron, I really appreciate it. We’d love to see you at BriForum in London, and we can make some special provisions for you. Otherwise, hopefully we’ll see you in Chicago, your hometown, and I see where you’re connecting from today.
Ron: Yep. Sitting out in the basement bunker.
Brian: Nice, well –
Gabe: And I’m sure that it is a bunker.
Ron: Yeah, it’s – I’ve got reinforced walls in my office, at least.
Brian: And room for your whole family. So on that note, Ron Oglesby, always a pleasure. Thank you so much for taking the time to join us today. Gabe, thank you for joining. Jack, thank you. And all of you, thank you for listening, and we will see you all – well hear – you can listen to us next week, same time, right here. Thanks a lot.