Q&A: Sun's CTO Talks About Cloud Computing

InformationWeek: In a recent keynote, you talked about the intersection of cloud computing and open source. What are the advantages for data centers?

Papadopoulos: The message there is, for those two things, the economic times will help accelerate their acceptance and not inhibit them. People are going to move rapidly to these types of environments. I think cloud computing and open source are deeply related to each other. If you're using a cloud -- especially a public cloud -- proprietary software licensing hasn't caught up there yet. It's really difficult to run Oracle or Windows in these cloud environments. Because the open source stuff is freely accessible, developers can make it happen more easily.

InformationWeek: So you feel cloud computing really is the Next Big Thing for the industry to focus on.

Papadopoulos: Well, we're all reluctant to call it the next big thing, maybe because we are all hype-cycled out. On one level this [cloud computing] is the path we have been on for a couple of decades with things like network computing. There are services that are presented through the network and that's a pattern we all know and understand and are comfortable with, so, nothing new there.

The thing that makes cloud computing distinct, where in fact there has been an evolution in network computing concepts, [is in] the developer on-ramp. It's different now than it was even two years ago. The stuff is actually accessible by grad students, startup developers, or people in IT shops. That's different. We know that when things are accessible and easy for developers, well, stuff happens.

InformationWeek: What can Sun do in bringing cloud computing and open source closer together that others can't do?

Papadopoulos: At Sun we do one thing -- we make systems. And the systems we make are network computing systems. Now, I'm not trying to oversell it, but the R&D portfolio I manage is centered on how to build scalable infrastructure, either in a public setting or an enterprise one. More than anyone else, Sun has placed an enormous bet on open source, and that's where you need developers. Virtually all of our software assets are there. We acquired MySQL largely to continue a connection with that community. I think it has paid off. GlassFish gets as many downloads a day as on MySQL, which is about 50,000 a day.

InformationWeek: So you're happy with the reception GlassFish has received.

Papadopoulos: One of the grand lessons we have learned in open source is, it is all about adoption. It is about what is accessible to developers, and what communities you build around it. For us, GlassFish represents a model community as we look at open source from a commercial interest. There's good participation among external people, there's a solid effort in the company to connect to that community. And there are concepts of different levels of engagements from casual users, to registered users, to being an active participant or contributor.

MySQL took a decade to build up to being this huge force in Web-based databases. And GlassFish looks like it is on the same trajectory to where it is the volume application server platform for developers.

So if you now think, well, having an application server as one of the cloud platform services, along with having relational databases and Web servers, you start to see this picture come together. These clouds are more than just the infrastructure-as-a-service-container model that Amazon has really been successful in promoting. We are making the big open source bet here.

InformationWeek: Do you have a strategic relationship with Amazon?

Papadopoulos: We do. You can get Solaris at Amazon -- that's a supported structure. I give them a lot of credit. When I was talking about making things accessible, Amazon, through great foresight or serendipity ended up with an equation that was simple enough so that developers could wrap their arms around it. They're saying: We're going to rent to you virtual machines and give you some storage, just bring your own stack.

InformationWeek: Are you surprised at the success they have had, being the only non-computer company you compete against in the cloud arena?

Papadopoulos: I think everyone is surprised by that. I felt like we were really early in with the Sun Grid. But what we didn't get that Amazon got was making it easy for developers to get to the on-ramp. At the other pole, we are looking at is what Microsoft is doing with Azure. I think they will be successful because they are going to just meet the needs of their own ecosystem, namely the enterprises in the SMB enterprise space.

InformationWeek: Will Sun invest in building huge datacenters, as Microsoft has done?

Papadopoulos: We're investing there, but we don't think that's the only answer. There will be these big public areas and you'll see more coming out on the Sun Cloud. We went with the Sun Grid, for instance, which is very developer-focused. So there will be these big public utility places for a GlassFish or MySQL developer. At the other end are enterprises that really want to have their own datacenters and control their own. Whether those will be co-located with these bigger ones, I'm not sure. There are lots of reasons why people will have private clouds.

InformationWeek: How enthusiastic are users you talk to about cloud computing? Are they ready to put some money behind a cloud computing strategy?

Papadopoulos: I can't get out of any customer conversation today without talking about cloud computing, open source, and open storage. This is really where the economy acts as this huge accelerant. It's like putting nitro in the fuel mixture, because all three of those areas have potential for reducing people's costs. The reason people may not have adopted them rapidly in the past has had to do with risk aversion. But right now with these economic conditions, you change your risk profile.

InformationWeek: Are companies adjusting IT budgets on the fly to accommodate more investments in cloud computing and open source?

Papadopoulos:

Yes, and in open storage as well. It works this way: If you're looking at your IT budget and you know you have to cut it, you look at what is discretionary and what isn't. At the top of the discretionary list is what you're doing with desktops and other departmental refreshes. The least discretionary spend is whatever you have with proprietary software licenses. If you stop paying Oracle, they tend to turn the technology off. Second is storage. At some point, the bits can't spill all over the floor, and you have regulatory or business reasons why you have to spend there. Then open source and open storage come up right after those conversations with a vengeance.

InformationWeek: Sun has a strong background in identity management. Are you able to use that as leverage to gain more technical support revenue and so a wider use of Sun's software?

Papadopoulos: If you look at, why did we go to open source several years ago with such a vengeance? Well, it's this concept of adoption. Developers don't buy things, they join things. They become members of communities and those communities, once you have a developer or admin in a community, then you get to have other conversations. Those conversations and the on-ramps will start in places like identity management, because everyone has that problem. The good thing is that identity management touches every system and so we get to talk about what's going in all these other places.

InformationWeek: Have you figured out where MySQL fits in with your cloud computing strategy?

Papadopoulos: Yes, very much so. There are two discussions there. There is MySQL as exists now as a relational database in Web and enterprise settings. There is also a relatively new project called Drizzle -- a fork of MySQL -- which is a very cloud-oriented so the name is appropriate I suppose. If you go to a place like Facebook and see how MySQL gets used, people only use a subset of the relational capabilities because it has a horizontal scale and there are all these concerns that go into large-scale deployments. So Drizzle sort of strips back down to a small core and then builds up the distributed capabilities.

InformationWeek: With Marten Mikos leaving, who had a strong vision for Sun's open source strategies, how big a hole does this create?

Papadopoulos: Marten is a great guy, but he is a builder. He didn't start MySQL but he made it into a company. I don't know if I would identify him as the visionary around open source, but he's very much a leader and visionary around exploring the business models that couple open source with commercial businesses. We will certainly miss his wisdom around that. But we have lots of ideas and lots of other prototypes of products for connecting to commercial stuff. For instance, how GlassFish has evolved in different ways. Virtual Box is another interesting one, along with the Java environment itself. For me one of the lasting contributions that MySQL culturally brought to Sun is the idea of making things very simple for developers to get going on. MySQL has this concept called the 15 Minute Rule. That rule says that from the time someone clicks download on the Web site to when they are running their first queries, it should take 15 minutes total. That's including download time and install. We have taken that philosophy and are driving it to other areas.

InformationWeek Analytics has published an independent analysis of the challenges around setting business priorities for next-gen Web applications. Download the report here (registration required).