 I'm John Furrier, I'm here at Google I.O. with John Schroeder, the CEO of MAPPA. Are you just in the session with Google announcing your big partnership? Tell us, what did you announce? Yeah, what we announced today, it's a really exciting announcement for us. So we're big supporters of the Google Compute Engine. We've been running on the Compute Engine for a number of months. Basically what we've done is worked out a partnership with Google, where we're the provider for our Hadoop distribution, we run big data analytics on top of the Google Compute Engine. And I think you'd tell from the session, people are pretty excited about it. You look at the cost comparison between running in your private cloud and the public cloud and saving money for people really makes them happy, right? So the Google guys are up there talking about the cloud. Obviously they run the big cloud, they have their Google. Amazon has a book business and a cloud as well. Those guys are smiling. They know a lot about the infrastructure and you guys are showing some stats up there. Obviously the cloud is benefit. It's like outsourcing. You can outsource a lot of costs. You guys showed a stat where you guys did a TerraSort, a high-end benchmark for data, big data. It cost you guys $16 to work it and the equivalent on-site deployed hardware. Hardware only was $5 million. And time to provision that was in months and you did it in minutes. Can you elaborate on that amazing stat? Yeah, if you look at what we compared is we ran a TerraSort on 1,256 servers and we were able to boot that up and run it in a matter of minutes. And we ran an entire TerraSort in a minute, 20 seconds. And yeah, if you look at the instance hour cost on that, it comes to about $16. The previous record or the record for TerraSort was run on-premise, on physical servers, on 1,460 servers with about four times as many cores and four times as many drives. And if you look at assembling that sort of environment, it had cost you $5,000,000,000 in capex alone and you'd have to have the electrician there, HVAC on the roof. It took you months to do, so it's just a dramatic difference. We've been following the infrastructure as a service business for a while. I've been real critical of it as a race to zero because it's like a hosting model. But what's happening in the big data space and applications in general is that the platform as a service layer has come in as a real differentiation opportunity where the pressure from applications is to actually have more power, more storage, more capability. Can you comment on what's happening in there? Because the speed factor and the cost is pretty compelling. But outside that, what are the other market forces that you see really kind of creating new life in infrastructure service and cloud services? Well, you know, I actually believe 10 years from now everything's going to run in the cloud. I mean, just a few months into starting my business over three years ago, I'd already put 20 tons of air conditioning on the roof just for my QA lab. And why do that? You want to sign your resources where they can move the needle forward on your business. And I think the general trend, while there's still a lot of talk about security and people's concerns about data security, the trend's in the direction of the cloud. I mean, we're putting all our Facebook information, all our personal information out there into Facebook. If you look at the amount of information you have in your Salesforce account, they know every single sales call, every deal you're working on, everybody who participated in those deals. Look at the information you put in Webex. If you look at your Webex report, it's every meeting you had, who attended, what the presentation was, maybe even a video. So while there's still a concern about security, certainly the trend is toward the cloud. And I think it's going to be a growing trend going forward. It's just more compelling than trying to build out these difficult to build data centers. So you guys have been in the news lately as we were at the Cube at Hadoop Summit and you guys were big supporters. I appreciate the sponsorship and underwriting there at the Cube. But you guys also announced a deal with Amazon. Yeah. Okay, now you've got Google. You've got kind of a thing with cloud. Why? The aligning with public cloud. Is there something that you're seeing that others aren't there? Yeah, I mean, there's pull from customers who want to run in a public cloud environment. So there's kind of two things that we get out of that. One is it's a very unique, compelling go-to-market, which is instead of going to a customer and saying, hey, let me sell you three years of capacity that you don't really need now, pay for the usage, pay as you go. And so it's a very low entry point for the customers and we're getting asked by large organizations to provide that service. If you're going to do that service, do it with the guys who really know how to build infrastructure. So it'd be crazy to build your own cloud service. You know, partner with Amazon, partner with Google, they're going to build out the infrastructure better than anyone else and then you can put your old world-class software on top of it. And then the final point I think it says who's really on top of the new market. I mean, we're all after these sorts of partnerships and you look at who Google and who Amazon chose to work with and it doesn't have far. I talk about the business model. Obviously, in the cloud, it's by as you go, it's like having electricity in your house and just a small little meter on the side rather than a substation, as I used to say to people. But let's talk about the business model. I mean, obviously, is it good for you to go by the drink and what's the environment because we said there's a pull from the customers? Talk about the business model around this. It's great for your business model if you're early enough to incorporate it into your business model. If your business model is predicated on getting people to buy all you can eat multi-year licenses or perpetual licenses and you need to pull that revenue forward, you can't do it in the cloud. But if you're early and you build that in, it becomes a wonderful deferred revenue stream for you, so it's really positive. And the way we do that also is through two different approaches. A lot, since we're such major partners, customers will go to AWS, they'll go to Google Compute Engine and they'll find us that way. So we can market through that from more of the consumer side of it, right? But then our enterprise sales force and enterprise sales forces of our partners, they can go out and sell blocks of big instance hours. And the deals get pretty big. I mean, a $300,000 instance hour purchase means we'll reserve that many hours for a customer. And there's kind of a couple of benefits there. One is we can approach that customer from a more strategic hands-on level that maybe Amazon or Google wouldn't want to do and help consult with them and make the choice. And then secondly, it gives the CIO the ability to manage their expenses because while paid by the drink sounds good, it wouldn't sound good if you were surprised by your cost per month going from 20,000 to 600,000 month a month and you didn't budget for that. So it gives them a way to say, okay, I bought, let's say, $500 an instance hours, I'm going to give it to a business unit, meter to that, that's what you're getting for the next three months. You know, public cloud when it first came out was great. Everyone's like, hey, put your credit card down, became a great place for shadow IT, great for developers. But when you talk to enterprise, if you guys are known for being the big data, enterprise, great, great, I do, the criticism was I'm not going to put mission-critical stuff on the cloud. I'll put do some black-shells analysis at night and I'll have to deploy some hardware. But in a way that is big data. So talk about what you've seen with big data, how that's changed the public cloud equation in terms of the kind of compute spinning up and doing the reserves. Because now you're seeing kind of that black-shells kind of analysis as a use case. Are the use cases expanding more with big data analytics? Are you seeing that? And is that the real driver here or is there something else going on? I think there's a general trend toward computing the cloud. I think big data analytics requires such scalable access. I mean, maybe you need 50 nodes today and you need 1200 nodes tomorrow. So I think that scalability is interesting. I think DR into the cloud is very interesting. So if you look at doing your on-prem and then you'd say, well, I want to do some mirroring to do DR in the cloud. I think that's another use case we see pretty consistently. So certainly quick scale up, scale down, and DR into the cloud are a couple good use cases. And so I think big data is a driver for it. But in general, I'm kind of a believer in the public cloud anyway. It's really early. So let's talk about open source. I see you guys are open source with Hadoop and you have some extensions. Is Google going to be open source? They mentioned on stage that open source. What's the aspect of open source? Are the open source and stuff that keeps them under a hard and top? Yeah, I mean, well, that's probably best for Google to answer. But what they said in there is the enabling tools are going to open source. So if you look at what's the value of open source, well, one of the values is ubiquity, right? So if you want people to build on top of App Engine or you want to build on top of Google Compute Engine and you give them some tools in the open source and then they can take advantage of those and you're going to get ubiquity out of that, right? They're certainly keeping the behind the data center stuff for the secret sauce, right? Like the questions that they responded to regarding what's your network infrastructure. They said, well, it's the really good Google stuff, right? Yeah, it's the stuff you can't see behind the curtain. Yeah. So it's a good enabler. I think that's some of the attributes of Hadoop as well. It's made an API or a set of APIs ubiquitous in the market and it gets an innovation there. Thanks so much for taking the time to speak with us. Congratulations. Great endorsement from Google. Essentially, the seminal paper of Matt produced well-documented with Google. You guys have some Google DNA. So congratulations on the deal and thanks for spending some time with us. Thanks. All right, great.