Q&A: BEA CTO Scott Dietzen

BEA Systems was one of the first companies to adopt Web services. Through its WebLogic platform, the San Jose, Calif.-based firm has even taken IBM down a peg or two.

Customers are clearly responding to BEA’s approach. According to a recent Gartner report, BEA’s WebLogic Portal has grown to become a leader in the enterprise portal server market.

However, since losing the title of “Java application server software” leader recently, the brains behind BEA have been focusing on its partners to pull the company out of its funk. The five-year with Big Blue was highlighted when BEA inked a deal with Hewlett-Packard as well as Salesforce.
com
on its new “sforce” initiative.

CTO Scott Dietzen has been at the forefront of the
battle, working on Internet application infrastructure for the past decade.
Dietzen came to BEA via the acquisition of WebLogic. Prior to that, he was
the principal technologist for Transarc. He holds a Ph.D. in computer science from Carnegie Mellon University.

Internetnews.com sat down with Dietzen to find out more about
BEA’s long-term strategy.

Q: These days so much is made about the importance of Web
services.What are you bringing to the table?

The overall mission is to turn the Web into an integration platform. Web
technology’s been very successful today, but the value proposition is all
about user interface whether it’s with data or a business application and
make it accessible. We, along with Microsoft and IBM, were the big winners
in Web application platform technology in the first generation. And as we
looked at how we could grow our value proposition, we said, “What is the
thorniest problem that developers and programmers have to deal with?” It’s
the complexity and ugliness of dealing with application integration tasks,
whether it’s behind the firewall EAI [Enterprise Application Integration]
or off to other businesses with B2B.

Q: What problems are you trying to solve?

The technologies to date have been hugely proprietary. They’re hard to
use they’re not well integrated. The typical scenario is you have one
transformation tool; one Web services tool, one workflow tool, one adaptor
tool and separate programming tools and user interface tools. So, the first
thing the integration developer has to do is learn all of these different
environments, build various piece parts and then try to figure out how to
integrate them together. The vendors haven’t even delivered on integrating
their own tool sets for integration let alone integrating what the customer
is trying to integrate. So the Web services value proposition is really
about extending the Web platform to enable integration. We’re commoditizing
a lot of the value of message brokering technology directly into the Web
container. The industry has started with XML, SOAP and WSDL and developers
are familiar with those. It’s the next-generation technologies that are
going to get us to the critical mass.

Q: To the critical mass? Which ones do you have in mind?

Critical mass for the Web was HTML, HTTP, SSL and URI. Those four have
created something much larger. I think in addition to XML Schema, SOAP and
WSDL, we now have Web Services Security. We’re shipping it. Microsoft is
shipping it for doing interoperability work between those two containers.
But the one I’m most excited about is Web Services Reliable messaging. This
is “application to application” messaging for Web services that gives you
guaranteed delivery for an order or a trade even as it crosses Internet
technologies. So the same quality of service you get for MQ Series or TIBCO
you will be able to get from Web Services.

There is also WS Addressing, which allows for asynchronous Web services
and WS Policy, which is a metadata model for capturing quality of service
or
security or other information about Web services — richer than UDDI. UDDI
has been one of the disappointments in the Web services stack in terms of
its adoption.

Q: Why?

I took an informal survey of our highest-end customers. We had our top
50
or top 40 end users in a user group together and I was astonished that all
of them claimed they had transactional Web services in production and 60 to
70 percent had transactional Web services running between .NET and WebLogic
in production. The VUDDI adoption was much less People were basically using
URIs and URLs as their model for locating and binding into Web services.

Q: What does that say about Web services as a whole?

It means we still have more work to do in the vendor community. I think
it’s also clear that UDDI has been oversold.

Q: Isn’t BEA working on a few of Web services standards of its
own?

At the higher layers of the Web services stack there is something called
XMLbeans that we’ve developed. It is not a standard yet but we are making
it
available via open source. It is a much easier way manipulating XML that
preserves loose coupling and gives you a very easy access model. If
somebody
changes the schema, the binding that you use will almost certainly continue
to work. Whereas, if you are groveling over a dom tree directly, that makes
for fragile programming. We all tout loose coupling as a benefit of Web
services, but the challenge is you don’t get this by accident; you have to
design for it.

Along the same lines and also just as exciting is something called XML
Query or X-Query for short. This is a W3C standard that BEA has been
driving
along with Microsoft and IBM. We’re convinced it will be SQL for XML. It’s
that level of ease of use. It’s much more efficient than XSLT and it should
replace the need for proprietary transformation and aggregation
technologies
in XML.

Q: It sounds like BEA, Microsoft and IBM are doing the lion’s
share
of the work.

We’re allied with Microsoft and IBM on the protocols and with IBM on the
programming. The three of us are doing the work around creating a standards
based platform for user interface, portal, integration and so on.

Q: Comment on your other partnerships — HP, Intel,
CapGemini, and especially Sun. Which would you rank as your top
collaborating partner in crime?

On the hardware side worldwide Intel, HP, Sun and Dell rates in the Top
10. Dell has done a lot of work replacing WebSphere with WebLogic in their
go-to market practice and their production environment as well, which has
been gratifying. Not surprisingly all of those vendors compete with IBM for
hardware and being a strong go-to market partner, we can offer them a
competitive leverage against WebSphere. That is one of the big motivators
in his partnership is this strong competition between IBM and us which ends
up elping our customers. The critical thing for us is that we’re quite a bit
smaller company than IBM or Microsoft. So our challenge is making sure we
get invited to the party. These partnerships are key because we don’t have
the kind of coverage that IBM has in the field. But when you add up our
systems integration and our hardware partners, we actually think we have
much broader coverage than IBM.

Q: BEA and IBM have gone back and forth as the leader in Java
distributions. What do you do to get in front?

We are quite prepared to concede that BEA and IBM have been winning
market share from the other server side Java platforms. No one really knows
in terms of revenues who is driving greater server side revenues between
BEA
and IBM. Everyone agrees we are winning market share relative to the
others,
but how our teams stack in the marketplace is murky. IBM’s financial
reporting is murky enough. WebSphere includes MQ Series. We don’t include
our Tuxedo revenues in WebLogic. It’s really an apples-to-turnips
comparison. This competition will continue. They seem quite prepared to see
us as their main competitor and co-market leader in several areas.

Q: With your development tools like WebLogic Workshop (now in
version 8.1), are you trying to simplify the process for less tech-savvy
people or are you making more things standard?

The reality is that there has always been five times as many business
programmers as there are system programmers. There are many more Cobalt, VB
power builders developers than there are CORBA, J2EE, distributed object
developers because business programming is easier than systems
programming — hard things possible, simple things simple. You have to have
both contracts. The business programmers do not want to learn distributed
objects and the distributed objects guys need the power of distributed
objects to do the things they do.

Q: Finally, Larry Ellison has said he is still looking to acquire
another software player beyond PeopleSoft. Can you comment on any plans the
company may have to merge with or be acquired by Oracle?

We don’t comment on any rumors and those rumors have been circulating for such a long time. We think the Web platform space is the best growth opportunity and BEA is the last pure-play in that space. We stand to grow with that market.

News Around the Web