Q&A: Sun Software CTO Hates 'Web 2.0'


At the DEMO 07 Conference in Palm Springs, Calif. last week, Sun's CTO for software Bob Brewin took time to talk about how Sun fits into the Web 2.0 world and the challenges the company faces moving forward. Brewin has been instrumental in Sun's strategy on developer technologies including the NetBeans IDE and Java Studio Creator. Lately he's been thinking about how to improving the developer experience, and alignment and integration of Sun's platforms, technologies and tools.

InformationWeek: What the most pressing problem that Web 2.0 faces?

Brewin: The name? That's one of the big ones. (Laughs) I'm going to use the term -- hate it -- but have nothing else to replace it with. One of the problems we have with Web 2.0 in general is that a lot of it is defined as all or nothing for a solution. If you take a look at REST-type [Representational State Transfer] services, these in many ways abandoned the tried and true technologies we build up in enterprise infrastructure, which I think are quite useful. So there's going to be challenges, for instance, in how we take enterprises, which are now moving into providing services for consumers or Web 2.0 portals that look into their business, [how do we take enterprises into this world ] when most of that they're doing is build around WS-STAR standards -- WS reliable messaging, security and so on. And most of what everybody else is looking for is not SOAP, they're looking for REST, right? So how do we bridge that gap and also handle things like security and identity in the new model? There're going to be areas like reliability -- these applications tend to be built fairly quickly. What happens when you have to scale out to a million users because your site suddenly becomes popular? What do you do about languages? What do you do about performance? There's a whole slew of different things that as we get into the space, and by "us" I mean the industry in general, there are a lot of problems that people just don't know they have yet.

InformationWeek: Is that largely theoretical or are people hitting these walls now?

Brewin: I think it's a mixture of both. The market, the business itself, is rapidly evolving. So there are some problems we just don't know we have yet in general. I've talked to a number of customers who are running into these sorts of problems. Granted, some of these are larger customers, so they have the engineering staff and financial wherewithal to go ahead and buy a thousand more systems. It's a problem for the small startup, without significant investment. Some of the companies... I hesitate to use the term "brick and mortar"... but they're old-style companies and now they're doing Web 2.0 stuff. They have a pretty good set of knowledge built up about how to handle scalability, how to handle security. [Startups], I don't think they've run into those problems yet but they probably will. And when they do, they're going to go look around for companies that have solved it or best practices or tools and technologies that can help them through their problems. I think that's an interesting opportunity for Sun.

InformationWeek: How is Sun approaching that? What does it bring to the table?

Brewin: I'll give you one example. There are many, but one that's the most tangible is in the area of Ruby on Rails. Sun recently hired the two guys who created the JRuby project, which is Ruby running atop the Java virtual machine. It has some huge advantages as a result. Number one, it brings at least a decade of worth of work being applied to the Java platform in terms of optimization, performance, and reliability. Building using JRuby means you'll actually be able to leverage that decade worth of experience. In addition, once you start using JRuby atop the VM [virtual machine], you can also utilize all the software libraries that come with the platform, as well as those provided by the rest of the Java community.

InformationWeek: Is Java more or less relevant today than it was ten years ago?

Brewin: If you're looking at it from a Web-centric, ultra-thin client model, Java is extremely strong in that sort of Web computing data center... so I think from that sense, the more Web there is, the more need there is for Java. If you look at it from the flip side and you say, "We've got all these different clients out there connecting to the Web," one of the true values of Java is that platform independence. Data centers are fairly consolidated, with only a few big players running this stuff. But on the desktop, there's a plethora of new clients coming out. So the ability to be able to run in a platform-independent manner atop all these platforms, especially as you start getting richer and richer clients, it actually becomes more relevant than it was ten years ago. Also, I think the platform itself is becoming more relevant. The Java language is one way of leveraging the Java platform. There are other ways. We can in fact, and it's been demonstrated in many situations, build other languages on top of the Java VM. So the language becomes less of an issue and the platform itself is increasingly relevant.

InformationWeek: Is Adobe's Apollo on your radar?

Brewin: Apollo is definitely on the radar. It's a very interesting technology. In many ways, it reminds me what we would have done in the past. But it's essentially the same thing. Java provides all of that already. I think the one thing that Adobe really brings to the floor is a rich set of tools that will allow content creators to leverage things like Flash and Flex. For the runtime experience, I don't see that there's any distinct advantage to it. In fact, I think there's a distinct disadvantage in the fact it's a proprietary, locked-into-a-vendor solution. In terms of addressing that market, Java can be competitive and probably provide additional value over what you can see in things like Flash and Flex. We actually do have a project that's under way -- we're planning on open sourcing it fairly soon -- it's called S Bridge. It is a scripting language built on top of Java. Hopefully we'll be doing a lot more talking about it in coming months.

InformationWeek: Is the consumer-driven push toward simplicity having an impact on enterprise development tools?

Brewin: Absolutely. I think this is trend that we've been seeing within Sun for a number of years. Project Rave (a.k.a. Java Studio Creator), that was one example of where we were trying to make the process of developing a Web application simple. I think if you take a look across what we've done with platform APIs or tools, the goal it to try to make it more accessible for more developers. Moving forward, what I think we need to do is take a look the content assembler space, [which is about creating a compelling Web or desktop presence from a collection of content]. For Sun, that means allowing people to create the infrastructure to host and stream media, to provide the tools to assemble that, the ability to take services developed by other companies and mash them up.

InformationWeek: Is that going to be tied to Sun's grid computing effort, as whole eco-system for application delivery?

Brewin: I think that's very viable and something we are looking at. For instance, one of the interesting problems that folks around here are taking a look it is if I'm trying to do a startup, do I have a lot of capital expenditures? Typically they go to hosting services. Hosting services today are primary organized around the Web but as these things become more complex, you need compute services as well. The things that are in the grid space definitely can help them. The utility computing model, I think, works. ...This pay as you go thing, I think, is the future. It's the model that works with subscription software. It's the model that works with utility computing.

InformationWeek: What are some of the challenges Sun faces at the moment?

Brewin: One of the more interesting ones for us specifically is visibility. Sun is a strong company and getting stronger...but most of the work people have been focused on at Sun has revolved around our core enterprise business and our systems business. Sun is doing a lot in the Web 2.0 space, the startup space -- we have a startup camp where we're basically providing equipment on the cheap to startups, almost next to nothing. There are all sorts of activities that Sun is participating in and the struggle is to make sure people are aware of it. It's the problem of where we show up at conferences like this and people say, "What is Sun doing here?"

InformationWeek: Does that entail having to play more in the consumer space, since that where most of the innovation seems to be happening?

Brewin: That's exactly a critical part of the strategy. In the old days, a lot of what customers received in terms of the Web experience or the desktop experience was driven by the enterprise. I think it has now flipped around where enterprises now have to adapt to meet the needs of consumers. Consumers want streaming videos. So guess what? That means we need to make sure that the networks, the computers, and everything else can actually provide those abilities.