Robert Brewin is a newly appointed co-CTO of software at Sun Microsystems. Tackling the application platform side of Sun software, Brewin is focusing on technologies such as Java, developer tools, Web services, and Web 2.0. The other co-CTO, Tim Marsland, keys in on technologies such as Solaris. InfoWorld Editor at Large Paul Krill spoke with Brewin this week about Java, Web 2.0, and what Sun is pondering in the area of software development.
Robert Brewin is a newly appointed co-CTO of software at Sun. Tackling the application platform side of Sun software, Brewin is focusing on technologies such as Java, developer tools, Web services, and Web 2.0. The other co-CTO, Tim Marsland, keys in on technologies such as Solaris. InfoWorld Editor at Large Paul Krill spoke with Brewin this week about Java, Web 2.0, and what Sun is pondering in the area of software development.
InfoWorld: So you have two CTOs for software?
Robert Brewin: Co-CTOs, that's why the "co" is in there. It's definitely a different kind of arrangement, and we do a lot of sharing. Both of us advise on both areas, and if you take a look at sort of the software platforms, the lines are blurring between what's an operating platform and what's an application. I mean a lot of the stuff you see on SOA and the Web 2.0 stuff, it's part of a platform, right? And if you want a network of a set of Web services that perform really well [and] are always available, it impacts on the operating system. And it impacts on the hardware. So it's all the way down to things like Niagara on a chip. Therefore, we do a lot of sharing. But if you had to pick sort of like the main domain, that would be it.
IW: What developments does Sun have planned in the area of Web services?
RB: We are doing two things. No. 1 is we're looking at completing and continuing development in the area of the WS-* stack, especially around the interoperability aspects with Microsoft. But in terms of when we start taking a look at the Web 2.0 space, REST [Representational State Transfer] starts looking more interesting.
IW: What are you going to be doing with REST?
RB: We're not sure yet. I'm working with folks like Tim Bray trying to figure out what the development community is looking for there.
IW: What's interesting about REST?
RB: I think it's primarily how lightweight it is. If you take a look at the WS-* stuff, it's fairly heavyweight, there's a lot in there to do a lot of sort of enterprise-class things. But if you don't need high levels of security or messaging reliability, then essentially simple Web services might be a better answer.
IW: I understand REST to be a more simplified version of Web services, which are not enterprise-ready?
RB: Yes, although a lot of enterprises I think are using it, again because it's simple. It comes down to if I'm going to be doing relatively simple Web transactions, whether I'm doing syndication or blogging, why do I need high levels of security? I don't care. Right? It doesn't really matter. If, on the other hand, I'm using a Web service to do a bank transaction, then it might make sense to use something like WS-Reliability.
IW: So what kind of products do you think you might release to support REST?
RB: Honestly I'm not sure yet. I've only been on the job now for about two weeks. So I'm hoping the next time we get together, I should have a lot more information.
IW: Which would be when?
RB: I'd probably have more to talk about within, let's say, four to six weeks.
IW: Is this some sort of development tool or product related to REST?
RB: Actually more generically, I would say it's more related to Sun's Web 2.0 strategy, and both in the sense of what we might be doing in developer tools as well as what we're going to do internally to support in terms of platform.
IW: So will there be a new product or a new developer tool?
RB: I think more than likely it will be changes to development tools, to better support Web 2.0 development.
IW: Related to REST?
RB: That is definitely one of the things on our list, the other ones being additional languages, scripting languages. [Including] obviously JavaScript. I'm looking at PHP [Hypertext Preprocessor ], Python, Ruby as possibilities. The other thing is to take a look at potential support for things like Ruby on Rails.
IW: Which tools are you talking about?
RB: I'm thinking of Sun Java Studio Creator and possibly NetBeans ... I'm thinking of Creator, simply because that's where we're doing a lot of our Web tier development.
IW: So this will be in the next few weeks and some kind of changes to the tools are being pondered. Do you know when these will be available?
RB: No, I don't yet.
IW: Don't you need to support AJAX [Asynchronous JavaScript and XML] to some degree?
RB: We support AJAX, and so we've done a couple of different things. One of them is obviously the Open AJAX announcement, right? ... Two weeks ago there were two announcements, we opened up two sites on Java.net. One's for JavaScript and one's for AJAX. And then the last one, there's an open source project called jMaki.
IW: How do you define Web 2.0?
RB: I like to think of it as a participatory Web. The problem with Web 2.0 is that nobody's ever been able to really pinpoint an exact definition. It's usually described as characteristics. But participatory seems to be one that works. Maybe interactive Web is a better way of putting it. .
IW: Interactive between who?
RB: Between the consumers of the Web and the people who are sort of the infrastructure. Or maybe the owners of the content.
IW: My next question here is about the future of Java. The Burton Group has talked about Java as kind of being on its way out. What do you think of the strength of Java?
RB: One of the things that's interesting about the [Burton Group] article is the article talks about [how] things like Java EE [Enterprise Edition] are being replaced by SOA and Web services, which I think it sort of funny because all of those things have to run on something. Those are just the way that you expose enterprise artifacts, whether they're data or services. And typically that's Java EE. I don't see how Java EE is going to go way any time soon any more than Java SE [Standard Edition ]. I think what you're going to see is probably a similar trend that you've seen with operating systems in the sense that it's the engine that runs the car, and so Java EE, Java SE are going to be a core piece of your infrastructure, which your developers are going to be talking to, or they're going to be talking Web services. They're going to be talking possibly different languages running on top of Java. So again, this is why we're looking at things like Ruby and PHP. But I don't think there's any danger of Java EE going away, because at some level the enterprise still requires the kind of scalability and reliability in things like transactions that only EE provides. The Web services themselves don't really do that.
IW: What about the complexity issue related to Java?
RB: I think we've addressed a lot of the complexity issues with Java EE 5, so I think that with EE 3 and EE 4, absolutely, we had problems. I mean that's why the launch [of] Java EE 5 with things like annotations just made it incredibly simple. So where before you had to have practically a doctorate in order to write an EJB [Enterprise JavaBeans ], now anybody who's a reasonably proficient programmer can write business objects.
IW: Is Sun making any money off of Java? We hear of Sun selling servers based on Java deployments and that sort of thing.
RB: It's more of a vehicle than anything else... It's sort of the derivative. There's two aspects to it. No. 1 is the derivative software that's built on top of Java, all the enterprise systems and so on that run on Java. Obviously there's licensing revenue for things like mobility, [with] J2ME [Java 2 Platform, Micro Edition]. And then the last piece is sort of everything that goes on once the software is built and deployed. So this is service, support, training. That's really where I believe the sort of future of software business goes. People aren't willing to pay to try out the software and build it, but they're willing to pay for service, support, and training after deployment. Look at SpikeSource as an example. There, you're taking open source stacks, you're certifying them, and then you're providing all the services to go with [them]. And I think that is a viable business opportunity and one that Sun's moving towards.
IW: I suspect we're hearing all this praise of open source from Oracle and Sun and IBM and everybody, but behind closed doors, I wonder if these companies are saying, What are going to do to stop this?
RB: To be fair, in some areas, sure. I mean that's sort of human nature. But what's interesting is that we can now point at things like GlassFish, right? The open source Java EE 5 server, it's getting more traction now than the proprietary one ever did. That's opening eyes... The interesting thing is that we are now recognizing direct revenue more than we ever did with the tools just by selling training on them... So it turns out that the theory is correct. If you make the software available and people build on it and deploy to it, there is a market for service for people that need 24/7 support. There is a market for people who are looking for indemnification from Sun, because we put our seal of approval on it. There is a market for people who are looking for training, whether it's developer training or sys admin training. Again, once that sort of demonstrable evidence starts appearing, a lot of the naysayers are just finally being converted.
IW: You also wanted to talk about application platform integration. What's happening there?
RB: This is mainly about the business integration space. This is the SOA space, specifically in the area [of] the SeeBeyond acquisition. This is business process modeling, business integration, service integration, tying that all together. My perspective is that that whole service orchestration and service integration business plays a big role in sort of Web 2.0 design.
IW: What's happening with SeeBeyond at this point?
RB: They're going to be releasing an update fairly soon. But most of the work is going on for what they call Java CAPS [Composite Application Platform Suite]. A lot of the work there is around building sort of the next-generation composite application suite with developer tools in an open source environment. So that's what's happening with SeeBeyond today. And then the third piece of that's related, and one of the reasons I'm going to Austin is trying to figure out how does identity and security work with that. Because I think the other piece that's missing in most of the current Web 2.0 artifacts out there is -- I want to make sure that my transactions are secure. How do I make sure that I am who I say I am?... And the other one is things like single sign-[on]. Think about a mash-up situation where I've got an application that's representing a number of different services which I may or may not have rights to. How can I just sign on once and have that federated across all of those service requests? Today, there's really no answer. I mean it's all sort of open and free and you're just using public services. But I can easily see where I'm going to create, let's say, a portal for me that I'm going to track my banking, my stock options, and all sorts of information that's private to me, in one federated portal using Web services. I'd like to be able to do that with a single password, single sign-on, and federate that across all the services. That's sort of missing today.
IW: What's been the difference in Sun since Jonathan Schwartz took over as CEO?
RB: Since Jonathan took over I think there's been a big change in terms of open source awareness and community. Having I think the only CEO that blogs is sort of an interesting position to be in. We have a CEO who is expressing his thoughts publicly in a forum, which is much broader than you see other people in his position [do]. The other interesting thing is if you see the kinds of things that he's talking about in his blogs, I think it encourages the rest of our organization to do a little bit more community outreach as well. So I think that's good. It also helps that he is the one who brought in Rich Green, who has also been a huge benefit so far.
IW: Sun has been increasingly moving toward offering hardware complying with the, for lack of a better word, Intel CPU standard. What does the future hold for Sun's own SPARC architecture?
RB: I think at this point I don't see any particular danger there. I think that what's interesting about the Intel architecture is that it's well-suited because of its low cost, to create a lot of low-cost boxes, especially in distributed environments and workstations. However, the latest advances in SPARC technology, especially around things like Niagara, are far and away superior to what they have in Intel for the particular purpose that they were designed for.
IW: Which is what?
RB: Which is, let's say, high-throughput Web transactions. Again, the kinds of things that run the Internet. A Niagara processor is more well-suited to run an Amazon.com or an eBay simply by throughput, the number of Web transactions you can support at any time. Because of the fact that it's got 32 cores on a given chip.
IW: Do you think that the SPARC really stands a chance when you have everybody thinking Intel?