BEA is a company on the move. And with its annual eWorld conference only weeks away, its promising big things for Java developers. Alfred Chuang, BEA Systems Inc.s affable chairman and CEO, spoke with eWEEK senior writer Darryl K. Taft last month about some of the issues surrounding the company. He also gave a few hints on what one might expect to see in the near term from the San Jose, Calif.-based Java infrastructure company.
How do you address the comment that BEA is just an application server company?
Well, I dont get offended by anything. If we were only an app server company and weve built a billion-dollar app server company, I think it is wonderful. This would be like when people were saying that Oracle is just a database company. Well, they built a $10 billion business, thats not bad. They have 45 percent operating profit; no one is bitching about that.
Its like saying Microsoft is a Windows company; it used to be a DOS company, and its made some of the richest people in the world. I think thats all cool. Infrastructure works. Everybody sees infrastructure software generate more profit than anything else—23 percent operating margins generate $300 million gross cash in a year. Thats not a bad business.
But the reality is when you look at what we reported last quarter, especially in the integration business, we generated $100 million in license revenue. Some of the other stuff is a little bit harder to specifically quantify, like portal, the impact of Workshop into selling—we sold 14 percent of our sales in the Platform SKU. And JRockit is the most highly proliferated JVM [Java virtual machine] out there in the marketplace.
We introduced a bunch of new products—Liquid Data, WebLogic Enterprise Security. So, I think the reality is weve been moving and expanding into the platform space and leading it way before anybody does it. So, if that is called an app server, Im OK with it.
IBM is your most direct competitor, and both you and IBM make claims about each other. BEA people say IBM fudges its reporting, and they say you charge more on the maintenance end. Whats the truth of the matter?
Well, I think the truth is something that nobody knows. This is the problem. I mean if you look at the most recent IBM report, their Software Group was the poorest performer in the company, which is supposed to be bringing in the most margins.
So, the way I look at it is they dont have to report actual numbers in terms of bookings and revenue on their distributed software. They never have to report that because theyre such a large conglomerate. So its difficult for me to say how we fare in terms of really competing with them. We have too, though.
We have to give out very specific numbers about this product category I have to certify for. So, this is really a certification contest more than anything. So, if they would step up and certify the numbers that they have actually generated, that would be very interesting to see.
The reality is this is a company thats been doing data processing for close to 100 years. Every one of our customers is, or they used to be, IBMs customers. So, we have to fight every day to bring these customers to us. And a lot of times, we have to convince the customers to switch.
So, these 15,000 customers that we have, I bet most of them if not all used to be IBM customers, or theyre still also IBMs customers. And IBM sells a lot of things to these people. I think thats just the reality of competition out there. Theyre a big conglomerate, and I think weve got to beat them by the merit and innovation in the marketplace, which I think we do pretty well.
Next Page: Expanding on the Workshop tool.
Expanding on Workshop
Theres your Workshop product, which has seen good reviews. Where do you see that going, particularly with Sun [Microsystems Inc.] about to deliver its Project Rave or Java Studio Creator this summer?
Well, I think our challenge with Workshop and where we stand is … Workshop is the first time people were able to see that its possible to do very complicated things pretty simply in the enterprise. We also showed for the first time the ability to include a very large group of developers with different skills in a company to do like things.
So, before we could never do it. This was only an opportunity on the desktop that people were able to do complicated things very simply. In the enterprise, it is totally different. Theres no simple thing.
So, I think the opportunity and also the challenge at hand is how do we continue to proliferate Workshop in an exponential way. So, we are thinking about several things to do. I think looking at continuing to share our technology with other people so other people can implement using our technology to adapt their tools or their back end of the systems.
We are looking at how do we reach a larger group of constituencies that tend to program to this stuff. Maybe a lower-end market, maybe expand the enterprise market itself that are already adopting other peoples runtimes that are very simple.
So, we are looking at a variety of different things that we can contribute. Some of them will be specs and some of them will be code that will give access for people to know all the underpinnings of Workshop, so they will feel very comfortable to build on it or below it.
And some of the things we obviously wont do, like porting Workshop to WebSphere. I think we will give them enough technology so that if theres demand, they should go do it. Youll see in the next few months that anybody can do anything to Workshop fundamentally from a systems software perspective, not only the end-user perspective.
So, I think then it will eliminate any possible concern that there is this level of abstraction in there from Java into this intermediacy for code generation. I think also, at the same time, itll be able to reach to a much larger crowd out there that currently is not using this kind of tool. Theyre using Emacs and writing code raw.
But the growth rate of Workshops adoption is very impressive.
Well, Ive heard different things about it. Some people have called you a mini-Microsoft, and that could be seen as praise. But with that comes the criticism that your tools are proprietary.
I think that concern will be totally eliminated within the next few months. And I think right now its an unfounded concern, but I understand why our competitors will use it against us. But I think once we do a few things in opening up specifically some source code, alongside with some very in-depth APIs, those concerns will all go away.
So you say wait and see?
What is BEAs strategy around service-oriented architecture (SOA)?
Theres no SOA platform today other than WebLogic Platform 8.1. This is the only SOA platform in the marketplace that truly—from development down to integration down to portal down to the actual runtime in application server, to the JVM, to WebLogic Enterprise Security—every piece of it is to be able to build, store, share, and structure and manage the whole services layer.
I think weve been pushing shared services and composite application model out there for quite some time.
I think this is the only real deal thats totally built specifically to do so out there in the marketplace. And I think we need to do more marketing, so that people can directly correlate WebLogic 8.1 into the only SOA platform out there in the marketplace.
I wouldnt want to challenge that as of today.
No, this is the only thing. The only other discussion is Microsofts Longhorn platform plus .Net, but even their integration story is very weak. And the other thing is every quarter they delay the thing two quarters. So, what are we talking about? Theres really nothing to compete with.
Next Page: An enterprise service bus for short-span messages.
Enterprise Service Bus
Well, there has been word from IBM, but its more of a services play and somewhat futuristic. But what are you doing in the area of an enterprise service bus (ESB)? Are you looking at that?
Yes, we have a project going inside the company that were not quite ready to discuss yet, specifically to address … And we do have a bus today, but there are several challenges. One is, in a one-to-many broadcasting model, most of the transactional buses are not designed for that.
Most of the transactional buses today are designed for very long-length transactions. And you have basically two-way communications between the sender and the subscriber.
The new model is more one where you have a significantly lopsided amount of broadcast messages that are small, short-spanned and one-to-many. And you have subscribers that will wind up being the only or among very few that actually take in the message itself. Right now, we have more of a many-to-many model.
So, we are in the process of building a bus specifically to handle these short-span messages. What has happened is the world in information systems is colliding with multimedia. All of a sudden, you have clips, you have pictures, you have things that people want to send, and they are waiting for reactions to them.
So, I think that has caused a different level of what people are expecting a transaction to be. What gets lumped into it? And thats caused us to rethink. Whats inside the bus that we have to do?
Any time frame on this enterprise service bus? This year?
Well, it looks like its going to be an early next year kind of time frame that well have something out.
I have to say that every time I talk to you guys, I get something new that I do not expect.
We have an endless appetite for doing new things. The other thing is WebLogic 8.1 just came out all done and is stabilized in the marketplace. So, 9.0—or whatever we will call it—is where our current mission is.
9.0 is a supplementary release, so its a nine-month release. So, everyone is now working on the plan for what will be in 10.0, which will be more than a year out.
And youve talked to Adam [Bosworth, BEAs chief architect and senior vice president] about mobility, we have a huge management project going on, the bus architecture edition will be a new thing for us, therell be a bunch of new things inside integration, as we are transforming integration even further and adhering more to our message of converging development and integration itself. Theres a lot of stuff cooking.
Thats a good point about the convergence of integration and development. However, one of the knocks against BEA is your integration story has been weak.
Well, we are now the No. 3 player. We knocked down webMethods [Inc.], with their $100 million license revenue. The next one will be TIBCO [Software Inc.] and then IBM, which they have a bunch of … I dont know what the hell they have, right—mostly mumbo jumbo.
But its all called WebSphere.
Well, its all called WebSphere or one of those things. I dont know what the hell they call it.
Can you talk about the mobility strategy?
One of the things we really want to do is make sure that from the development and deployment perspective, applications are ready for mobility right away.
Our vision is the following: Right now, you look at most of the decision-making applications and they are still very much fixed to the desktop, despite that there are portals and other things out there. Its a nonmobile kind of application.
Most of the decision-making in companies is very short-spanned. Do you approve this job offer? Yes, click. Its a very inefficient process. You see this via e-mail, which is how most of the things are initiated. When you get e-mail, you click on the URL, you go inside the portal, you still have to fill in a bunch of stuff and approve … totally useless.
Now, how do you turn a single application that can go from this cumbersome process into a very simple set of questions with some short-span questions on it unless you have a little device that can say, Ah, I understand. OK, this makes sense, Im going to approve.
And I think that is the design goal. How do you make that available? You sit down and start building the application. These apps are deployable and manageable right off the bat on mobility devices—so it really is an enabler.
Next Page: Chuang says none of BEAs customers is looking for Eclipse.
Do you think that BEA is at a disadvantage because youre not involved in Eclipse?
Well, heres the thing. We are a customer-driven company. If our customers come to us and say, Well, you have to do Eclipse, otherwise I wont buy your runtime and you are inhibiting me from doing what I need to do, well do it. So far? We dont have that request. And on top of that, I think people are barely digesting the real deal, which is Workshop out there.
Standards and a set of frameworks to me are useless unless it is being applied. I think its much more practical when you have a full set of tools and the framework already done, and you allow people to adopt what is in there from a standard fashion. And I think our approach is a much better approach than doing Eclipse, which none of our customers is looking for.
A lot of marketing went into the program. Marketing doesnt mean anything in our world. Look at how many programs IBM has gone through around this stuff, around component building. Oh my God. There was Project San Francisco.
That kind of reminds me of a question I put to another BEA executive about your support for J2EE 1.4, and his response was, well, if you want Web services support, we already have support for that, so we dont need to necessarily adhere to J2EE 1.4 right away to give our customers Web services support. So, the question is, when it comes to certain standards, how do you decide which ones to jump right on and adopt?
I think we will always support the standards. The problem is … lets talk about the people in engineering. We have 700 people in engineering. There are so many interesting things to do—inside and outside the standard. So, the only thing that makes sense is if our customers are saying they want J2EE 1.4 right away. Then well do it. If theyre not asking for it, we want to put the resources elsewhere that would really benefit them—which is really what were trying to accomplish.
So, youll support it in the next version?
Absolutely. But you dont have to rush to it, knowing that nobody is going to adopt it right away. Many people are barely getting into 1.2. Academic exercises in the real, commercial world sometimes dont apply.
I asked you this before, but now its a year later, and IBM has digested Rational [Software Corp.]. Do you see that acquisition to have made the competition with them any different?
No. I think its far less. We have never heard of this word Rational anymore, even in our large accounts—and I think that is just reality; its amazing.
I always pray the things we dont want, our competitors will buy them. Then, they generally do a poor job of integrating them and they get put out of the market.
Well, with the modeling technology theyve added to their lineup, you dont see that as an advantage for them?
Well, UML [Unified Modeling Language] is interesting technology—its an intermediate language scheme. And its huge. You look at most of the UML models, you say, oh my God. They never generate any code out of the things. Its just modeling for modelings sake. So, for me, thats not very interesting.
There are much more interesting things to do. If this was the kind of new technology people are doing around BPM [business process management] workflow, that is much more interesting. It captures specific business flows and allows you to dynamically change them on the fly.
This is so much more usable than modeling their entire business and trying to generate code to run it, and you always have to go back to the model and adjust the model itself. I dont think thats the practical way to go anymore, with the technology we have today.
Next Page: A critical look at Java governance.
Eyeing Java Governance
So, are you saying you wont support modeling in your toolset?
Well, there is the ability today, you can abstract and you can import the [Rational] Rose model already. But to support it meaning we will support UML? Our customers are not asking for it.
People are looking for simplicity. They want to have loose coupling. They want to have smaller blocks of shared services applications, not to have one giant service that encompasses the whole thing. Its just not what people are chasing anymore.
Do you have any stance on the whole open-source Java debate?
Nobodys happy with the way that the Java governance works within Sun today. Ill point out some interesting things. Its a taboo in our world to name anything Enterprise Java or Java X in your product.
You look at what Sun now calls its Java Enterprise Systems—all the LDAP [Lightweight Directory Access Protocol] services, none of those things are written in Java; theyre all written in C. Thats just not right.
So, I think from that perspective, the governance needs to be revisited. I dont know if open source is the way to solve that problem. Because one of the things that we enjoy today is the reason why Java has progressed so fast is because of the lack of bureaucracy. When you have too many cooks in the kitchen, it stalls the progression.
Now, can open source potentially solve that problem? If its well-run, it can. But who is going to run it? It is an enormous investment to run open-source efforts like this. It would be bigger than what Sun is putting into just governing Java and doing reference implementations.
I would have a hard time with Sun switching from that into running a big open-source organization. I just dont see it. I dont know what the benefit is anymore out of this whole Java thing.
How important is the government market to your company?
Its hugely important. But I think most of the time people are only looking at the U.S. government, because thats where a lot of the spending has been. Our government business spans around the world. I would say most of the tax agencies use our software for their online reporting.
Our customers include postal agencies, a lot of military deployment for intelligence capturing and military control logistical management systems in and outside of this country. And a lot of the government-to-citizen interaction portals are developed on BEA.
Our government business is about the same inside and outside the U.S. This is a critical market for us.