February 10, 2016

New York

Linux on the Laptops, Docker in the Labs, and Migrating Monoliths

So as Brian mentioned we're gonna talk about Linux on Laptops, Docker Labs, Migrating Monoliths and I have to say I just kick it off, probably what a lot of people are thinking is that when we talk about things like Linux laptops Docker labs, Dell CTO's office is not maybe the first thing we expect. You wanna talk a little bit about what you're doing at Dell and how this all makes sense? >> Yeah.

I think that that's a good introduction. So I am Barton George. I am with the CTO office in Dell. Been at Dell for about six years. A couple of different jobs. Started as the cloud evangelist. What we do in the CTO's office (which I've been in now for a little over six months) is we look at technologies probably about two years out.

So it's not the real far gazing ten years down the road but a lot of the folks in the group what they're doing is looking at standards looking at what's gonna happen, and as you would probably guess a lot of this is around hardware and around servers and what technologies are gonna benefit customers the best down the road.

That being said, we're also looking at how do we differentiate Dell going forward beyond hardware? That's our bread and butter but at the same time we need to let people know that we're thinking about other things because we are. And things that we think frankly matter. Things like the Linux desktop, things like Docker, how do we connect the laptop to the cloud? And that's basically what we're here to chat about.

>> Let's talk a little bit about the Linux laptop or the Linux desktop. As Bryan just mentioned there seems to be a little bit disturbance in the force and you're starting to see, Linux laptops pop up here and there more than perhaps they have in the past. Now you've been doing this for a while though, you hatched this idea back in 2012? >> 2012 Yeah, four years ago.

>> Yeah , you wanna take us back to what started this, what provoked it? >> So a good question so, I was in- >> My job. >> Exactly. I was in the group between the cloud group and the CTO office, I was in a vertical called the web vertical, so you had the financial guys, you had the retail guys, you had the oil and gas, the group that I was in, what we focused on was a group of two, we focused on people like the Facebooks, the Googles all the way to the startups, and so we had brought an analyst I don't know if anyone knows Stephen O'grady of Red Monk, there a boutique firm that focuses on developers and were trying to figure out hey if we're in this web space really the developer is king here.

They're the people who are leading the way they have the most sway particularly in this audience, now when I say about the web audience, it's true for others, as well- >> This audience and I've been dying to ask, in this audience is anybody using a Linux laptop? Anybody raise their hand? >> Wow! Awesome.>> How about that? >> Wow I thought it was going to be a Mac attack.

>> All right, so what we did then was he came up and he said, you know what, no major OEM has got Linux running on a laptop pre-loaded, if you guys can do that, I think that would really be good, there would be a market for it, and then I thought, hey, great idea, it's never gonna happen, just because at Dell we do lots of volume and this would not be in the realm, this would be like what we're doing in Belgium on Tuesday between two and four in the afternoon so it was gonna be really tough to this make sale, so I just sort of filed it under away, good ideas that will never happen.

And then this is where the internal incubation fund came along. So Michael Dell's executive assistant put together this thing, and said, hey people around Dell if you have an idea, come and pitch to us we'll be an internal angel fund, if we like your idea we'll give you a little pot of money and you get six months to try and prove it out.

And so that's where I pitched and I said hey guys I think it would be pretty cool if we did this, Ubuntu based developer laptop. And so I got the okay, and then as I said a very small pot of money which they ended up only giving me half of it anyway, and away we went. >> Okay. Well so, one thing that struck me from some of the sessions here, earlier today, Matt from the Hudson Bay Company talked about, the technology challenges, and also the cultural challenges, having to sell his initiative to sell microservices Docker into the bank or I am sorry into Hudson's Bay Company so what struck me is that it's a technical challenge but a cultural challenge, and we're not talking about bring microservices into an organization, but you brought a Linux laptop to life in Dell in 2012, fast forward now is fifth generation.

>> Right. >> Fifth generation coming out this month. Put your orders in now. >> So can you talk a little bit about. How did you get there? What's the history? How did you? >> Yeah and I think. So as I said, it was very much a skunkworks project. We had a bunch of us who were working part time on this.

If we had sort of a 20% program at Google, that would be what this was. And so what we did was the idea was let's do—this is how I pitched it to begin with. I said let's do this all out in the open. Let's be transparent. We'll do it all in my blog. Telling people what we're doing. That being said, before we did that, we did the Let's Not Be Stupid phase, so we went first and got three alpha cosmonauts in Project Sputnik right.

So there's this Russian space theme going. >> And why is it called Sputnik? Already asked that question. >> Because Mark Shuttleworth who founded Canonical was the second space tourist. He went up in Soyuz but that's not as cool as Sputnik. So they went with Sputnik. >> Oh okay. >> So with this we have three alpha cosmonauts, I know Dustin's gonna be speaking at some point.

I don't know if he is in the audience, but he's Dustin Kirkland, he was one of the three cosmonauts. So we just bounced the ideas off of them, next step is we took it out to two real big web properties on the West Coast, and they didn't come up and say we'll buy thousands of them, but they said that looks interesting, come back when you're a little further along.

And so we realized, okay, we're good to go and we launched on my blog. And basically when we launched, we didn't even say here is a product. We said, here is something that we would really like to get going, we really want your input. We don't know if this is ever going to become real or not.

But if we get enough positive input, we'll make this a real boy. So just to give you an idea of what the traction was in my blog, I think leading up to it I had hits of during the day of one day count was 60, one was 70 one was 97. Then when it hit it went up to 5000, 8000, 15,000, and now I've had over- >> That's like 15 minutes in Belgium? [LAUGH] >> Exactly. And then we had over 70,000 views on that now, but the big thing is that we asked for feedback.

We got tons of feedback and it was because of this that we ended up saying—the tipping point was when we said okay we're now gonna move to beta. And we asked people who want to be beta customers. We had 6000 people around the world and that's when we said, okay, This is going to become real.

Let's push this forward >> It was just that easy >> You onto the fifth generation— >> No no no. So then there's the antibodies and I think we talked about this a lot in other companies here during the session. I think it was Larry from the Hudson Bay Company talking about they've got other things to do.

You think you got this amazing idea that if they just listen that it would just be awesome. You've gotta remember that everybody else has got all this projects, they're already overloaded and you're looking to take cycles from them because this is an outer scope process. And so there constantly was roadblocks that we would come up against And you feel a little punch drunk by the time you finally go and launch, and I think we were talking beforehand is it doesn't stop when you launch your first generation.

It would be every sort of few months after that. Somebody new would get wind of this and say why the heck are we doing this. You're taking resources away from valuable projects and we'd have to do the fight all over again. You have to be eternally vigilant when the zombies come. >> So tell me one more thing that sort of looking back on the history that's well chronicled on your blog is your original vision was bigger than just a Linux laptop.

Could you say a little bit more about that, and what that was, and what it ended up being. >> Yeah, and I can't remember who was just talking about this? I think it was the gentleman from Uber, but one of the ideas is hey you got a new language that's kind of a pain to set up, so we thought and we took a page out of pivotal labs who had this type of a setup for Macs which is basically language docs that you could pull down, and you could set up on your laptop so that when you started somebody new, when you switched systems, you could easily set it up.

So we wanted to do what we called was this profile tool and we had a big idea how people in the communities were gonna build it, put it up on GitHub, you're gonna be able to pull it down. So that was one vector. The other vector was what we called the cloud launcher, and it's the idea of let's really connect this to the cloud and what we'll do is—remember this is 2012 right—so we'll develop these micro clouds on the laptop we'll use LXC to then push it up to the cloud, we use that in juju.

Those turned out to be really tricky so we put the cloud launcher aside, we focused on the profile tool, and that became our Vietnam, right, it was a quagmire, we kept taking run after run at this and we said this time we are really serious we really have the resources we are going to do it and then it never happened and finally we had to give up.

>> So you gave up but serendipity. >> Yes serendipity came into play. So this thing called Docker came along, which has come up a couple of times at some of the talks, and the idea was wow this is really what we were trying to do. We were trying to reinvent the wheel and then the wheel came to us. So you now have this ability to develop natively on your laptop, you don't have to use virtual box or use VMs as you might use in other Operating Systems, and you can create it there and you can push that same application directly to the cloud so as I said we sort of backed into this and got lucky. It wasn't our original method of getting there but history we are on the right side of history I guess or something like that.

>> Before we move to some of these other topics any sort of lessons learned that you think might be applicable to folks that maybe are trying to bring Docker into their organization and encountering some of the same kind of antibodies that you might have encountered? >> Well I don't know if it's Docker per say, but in hearing people's talking about bringing Docker I think the lessons we learned are very similar to what you have introducing any kind of new technology or a new way of doing things. And I think one of the things is to get a high level champion, while at the same time you yourself need to be vigilant.

You need to always sleep with one eye open for these types of things, and you always need to be pushing them. I think another thing which is a theme I've heard here is, you start small, you start it a pilot. We probably shouldn't have marked with the profile tool and the cloud launcher to begin with, because that became a bit of a distraction, but one thing we did do was just start with one system and one config and we've grown that from there. And then if you doing this out in the open it's really about straight talk, admitting it when you screw up cuz it's not are you going to screw up, it's when you do screw up how are you going to recover.

So just being open, being transparent, and I think and iterative which is very much a lot of the things that we've been hearing. I mean it's basically bringing an agile process to product development and leveraging the power of the community. >> So if anybody is followed Barton, on Barton's blog is that it's called? >> Bartongeorge.net >> Bartongeorge.net you know that he backs this up, definitely open that is what you are doing and you are also a bit of, I don't know, maybe a serial entrepreneur within Dell is a way to think about it? >> Intrepreneur.

>> Intrepreneur. I wanna talk about Docker labs a little bit, because what you've been talking and blogging and posting videos about lately has been about the Docker lab that you've been building. So can you say a little bit about that what it is and why you are doing…? >> And it's another one of these approaches—I gotta be careful how I position this—but we as a company had looked at a container strategy, and as you can imagine it ended up being one of these huge PowerPoint presentations it sort of collapsed under it's own weight nobody ever read.

So what we did—and the great thing is I work for the CTO, and he is very much a pragmatist. And a crawl walk run type of approach guy. And so that was where we said if we are gonna do container let's not put together as grandiose plan. It would have been awesome if we had a spare billion dollars but we don't.

So what we did so let's start small yeah exactly let's start small and let's get a feel for what we got and learn from that. And that's what we've set up in the lab we've got two nodes, CTO office has our own lab future bill one and future bill two and we've just set up nodes, head note and compute node got doctor running on it.

We are using Triton. And the idea their is let's start learning and figuring out from here and at the same time. While this isn't open. Open for people to access publically we are talking about it publically and the whole thing is works share our learning with folks. >> What's the one thing that you've done different than many that I see are sort of dipping their toes into the pool is.

You haven't just crabbed just a couple of VM's on a public cloud or thrown up a couple of VM's on your infrastructure. You've started from the. The network up can you say a little bit more about that. Why did you take that approach? >> I think what we want is as a company we are looking for the full stack, did doesn't mean that everyone is gonna always buy the full stack from us but we wanna at least understand how it all plays together.

And I think that's something you loose from going up the cloud. At least what. We would loose so we do we have storage, we have networking, we've got the nodes themselves and it's how do they all work together. Now we would like to connect it to the cloud same API's but that would be in a second step and right now it's just as I say it's learning for us and try to figure out how it functions.

I think that that's important, important contribution that you're making. The CNCF panel that was here earlier, if you guys remember that, one of the things that was highlighted as contentious issue was how we're gonna network these 500 servers that Intel is coming to the table with.

So being able to understand the full stack as people are moving in to the private Private cloud as well as public cloud with these solutions and the most surprising thing for me out of the key note this morning from Forester was that although the pronouncement of the private club being dead is certainly.

>> Greatly exaggerated. >> Maybe a bit exaggerated and there certainly some of that still to be done. How about just in terms of how it's helped you in your team, do you have any. >> I think one of the greatest things is we talked a lot about culture really is a way to get our folks within the CTL office little more comfortable with the idea.

Let's get in there and iterate. We tend to do secrete things that know we very well and we're huge experts on and that kind of way you have a lot of control you can do things methodically, you can work as you're used I think with this is that areas that we're not comfortable with and so the whole idea of getting in there One of the guys on our team whom you worked with was very much in the beginning what's our container strategy? What's our huge vision for this and didn't really want to get into the water until we had the whole road maps sketched out.

And happy to say he's come over to the good side and And now was comfortable but just getting in there >> Looked happier when he was able to put his hands on - >> Yeah. Exactly. You know it goes from being PowerPoint talks to actually being getting your hands dirty and that's the kind of thing.

And I think that's one of the thing that when I was hired into the CTO office was something they were hoping I would bring Through my experience through project [UNKNOWN] Makers is just getting in there. And once again do get in actual kind of a way, doing it openly, not having to have the huge path all the way but to try and learn as we go.

>> So one more panel question that I wanted to introduce here, so who in the audience still has a flip phone ? [CROSSTALK] Somehow I cut out there. Flip phone, anybody? Barndon/g do you wanna confess? >> Yeah so one of the things I love doing is I come to conferences like this, like I would say [UNKNOWN] I think I did 10 interviews I just go and meet people from Talk to people from Corel masses/g here etc, etc, and what they're doing.

Brian and Scot were a couple of victims that I did. But yes I do use a flip cam - >> So why are you bringing that up? It's certainly not too, although it might be retro cool now - >> Yes it is playing pong instead of whatever I have seen and you've told stories of how really effective these videos are, they just says, not getting too hang up on over producing.

>> Yeah I mean there are gorilla style videos I love them and- >> I mean internal and external. >> Yeah, and there is no dell connection with these two which is one thing that I try to do is I don't Cuz I don't, no one comes in and says we love your servers, but it really helps, I go to a conference and then help them [UNKNOWN] Just plug it in, upload it and blog it, just, once again the agile method it's not about getting the amazing set up the camera crew Through and get the lighting and spend five days and edit it e.t.c.

It's all about what can you get out there now, what can you, even know that it's not perfect and that adds to legitimacy of this and as I said it's expediency and it's the newness of all. >> So watching what you've done in your blogs and with that I think the reason why I Brought it up is important to be able to sell, right internalize projects.

So last couple minutes I wanted to ask you about what's up here is migrated modulus. And so you have the starter lab and we see that a number of folks that we work with are starting there. I think it's a great way to get started, get your hands on something. But certainly one of the pieces of advice I give the teams that we work with is that this is not appeal to dreams this is not if you build it they will come.

There's a collaterally that if you don't build it they will go right?>> Right >> Because they wanna have a place to leverage and build out their micro service or architecture and deploy their Docker solutions but it's important to also think Think about doing something with lab.

So we can talk a little bit about how to give approach to that from an application perspective >> Yeah I think with a lot of these things in big companies I'm assuming a lot here are probably from big companies you find pockets of quote and quote enlightenment people who're just doing things on their own.

One guy had taken, he wanted to learn about containers and so So he just took one of our old school apps and started to containerize it. And I think that the one that you're referring to is we have something called ASM application system manager, which basically manages systems etc.

>> Short name for a monolith ? >> Yes exactly. And so we're taking that now and the first thing is just throw it all Into one container, that's step one. Once again, it's the baby steps. Next step after that is how do we break that up into micro services? And so it's the idea of what we'll get more agility in the end, it'll be better for the customers but it's also learning from for us and we want to share that learning cuz there's probably other people like Dell Who are large companies who have applications that they're trying to figure out, how do we take that and put it into micro services.

Let people learn from our mistakes and some of our successes as well. >> So we've covered not a lot of ground and we already out of time and to make sure you guys get onto the break and onto the next sessions. Thank you for that. If you are interested in following along with a carton/g and older seemer doing it is as he said, all out there, all out in the open, the good the bad and the ugly, of course you can look at Buttongeorge.net, - >> Yeah buttongeorge.net >> and be weary he's got a sick phone in his pocket he' ll Shoot a video of you so, thanks Button.

>> Cool. Thanks everybody. >> [APPLAUSE] [BLANK_AUDIO]


Barton George: Senior Principal Engineer, Dell

Bill Fine: VP, Product and Marketing, Joyent