CRN Interview: Mark Carges and Scott Dietzen, Bea Systems

As a company, BEA spends each day battling IBM to be the dominant provider of Java-based application server platforms. The next step in that campaign is to establish its WebLogic offering as the premier platform for integration. In an interview with CRN Editor in Chief Michael Vizard, Mark Carges, president of BEA's Enterprise Framework Division, and company CTO Scott Dietzen tout BEA reliance on channel partners to compete with IBM and explain how Web services technologies are changing the economics of integration at IBM's expense.

CRN : Given that IBM is your primary rival, how do you contrast yourselves?

Mark Carges (MC) : Our agenda is to provide a unified, extensible platform that customers can build applications on very rapidly and cost effectively, as opposed to IBM, which has a portfolio of technologies, and IBM Global Services will get you just the right thing to put it together, which is very labor intensive and not cost effective.

Scott Dietzen (SD) : We need to put together a best-of-breed alternative that is cheaper and has more functionality. The answer for us is to leverage all our partners to challenge IBM's one-stop shopping. Last quarter, we won 222 hands-on engagements against WebSphere. We desperately need our channel to make that happen. The real challenge involves the dollars needed to add value to the platform in order to get to the solution. We hear IBM is anywhere from $11 to $18 of services on top of $1 of software license revenue in the WebSphere space. We're about 30 percent underneath that. What we see IBM doing is discounting the software to compete with us and then tying in the services deal. That's why it's crucial for us to go to market with Accenture, EDS, Bearing Point and all the rest so we have a good counter strategy.

CRN : Why should a solution provider go with BEA if there is more services revenue available working with IBM?

SD : The reality is that a solution provider can go in and bid a deal and deliver a lot more to the customer. There's not a shortage of IT problems that need to be solved. They can either underbid IBM Global Services or out-deliver them. With us, you have either option to win the business. Another issue is that at some point these costs are going to be too big for IBM to maintain and for the customers to maintain, so customers are going to push back. There is going to have to be a dramatic simplification and historically, this is not something that IBM has done well. Ninety percent of the stuff inside WebSphere doesn't live anymore after five years because it's all legacy stuff.

CRN : How will that simplification manifest itself?

SD : At the end of the day, if the Java development community doesn't get ease of development right, Microsoft will and IBM and BEA will be relegated to the margins. Ease of development is probably the single biggest way to differentiate against IBM, outside of cost and or collaboration with Intel and Accenture. A big part of this is that we need to build better run-time models behind the tools. The power of a Visual Basic or a PowerBuilder is the run-time architecture. That's what our [WebLogic Workshop initiative has been all about. It's overlaying a simpler run-time model on top of J2EE. Another big issue is tying the various tools together. With IBM, a developer has to pick up and learn five or six different environments to do an integration task. That's what makes integration such an un-fun job today. The integration products themselves are not integrated. BEA has been just as guilty as any other vendor. We're all going to market with software for integration that isn't integrated. But we're moving to a model where everything gets surfaced through Workshop.

CRN : Speaking of tools, BEA launched a data-integration offering called Liquid Data. Where does that product fit in the scheme of an overall integration strategy?

MC : The two key technologies in Liquid Data are XML and XQuery. The first thing we do is put an XML schema on top of the data sources. If you give it a data source with some kind of structure, we can put a schema on top of it. The beauty is that you now have a single way of looking at all the data sources as a set of schemas. From there, you can build a logical view of the data and you have no idea if the data comes from SAP or a mainframe. From there, we can generate a query that is built using XQuery.

CRN : What is XQuery?

MC : It's a brand new standard that's still being crafted in the W3C. You have Oracle, IBM, Microsoft and us all very interested in it. There are things that it can do for data access and transformation that other things couldn't do very well. It has a sweet spot with data architects because it's enough like SQL, but it's more powerful. This is the first implementation of XQuery in a larger data solution.

CRN : What's the ultimate benefit here compared to existing approaches?

MC : Liquid Data is about real-time visibility into the data. The key thing is visibility. You could make data visible without Liquid Data. You could use traditional EAI approaches, use adapters and write serial processes to make things visible. Other companies have tried to solve this by federating data into SQL, but that ties you to the repository. BEA is independent of the database or the application. What our customers really like about that is that we're not trying to sell them a database. We help them virtualize all their data sources so they can query them.

CRN : How does the emergence of the application server as an integration platform change BEA's relationship with traditional EAI vendors?

MC : We're now a better integration partner than say a WebMethods, IBM or others. Our solutions provide visibility and process, where their solutions only deal with process. That's really our sweet spot. We're out to solve integration problems that include both data and process. In the overall integration space, we compete with them. And over time, I think the customer will see getting everything from one vendor will have a lot of economic advantages.

SD : These guys are not going to be able to keep pace with all infrastructure they need to plug into a Web integration platform. It becomes a game for the big guys. It's just way too expensive and it doesn't make sense for the industry to develop it 10 times over. It's going to take critical mass and most of these integration companies will find they won't be able to stay in business.

CRN : How would you describe BEA's positioning as it relates to the big providers of enterprise applications?

SD : In terms of integration, PeopleSoft and Siebel are being a lot more aggressive because they are finding that their chief barrier to sales is integration. Customers are not buying until they understand how an application ties into all their other applications. These guys are working very aggressively with us and our competitors to build a better integration story. SAP and Oracle see the world differently. Part of their strategy is to say that integration is too hard and that's why you should buy all your software from them. It's a worst-of-breed strategy as opposed to best-of-breed strategy. But the trend is sharply in the other direction. Software is getting more vertically specialized and there's no chance that a Merrill Lynch could run their entire business on top of just SAP or Oracle. What we've got to do is turn the Web into a fabric that can tie those suites together. Hopefully, that is what all the work around XML and Web services is about. It turns the Web into an integration platform.

CRN : Given all this, why does it appear that IBM is making headway recruiting ISVs to the WebSphere platform?

SD : Privately, IBM Global Services will tell ISVs that they will not go to market with them unless they support WebSphere. The larger ISVs are intent on supporting WebSphere and WebLogic. IBM warrants they have a WebSphere story and IBM has leverage that is in their interest to exercise.

CRN : How does BEA feel about the open-source movement?

SD : With open source, for the first time there is a community that has a delivery vehicle that is cheaper than Microsoft. We're doing a significant part of our business upselling more security, more management, more reliability and more performance into open-source deployments. It's a great bridge to get people started on Java. But there's no comprehensive open-source J2EE container. And it's not in IBM interest to drive this. Who's going to drive a full J2EE container into open source? It's not going to be us, IBM or Oracle. And people don't look to Sun for cross-platform software.

CRN : What implications does the advent of XML have for database vendors?

SD : One of the things that the database guys are disgruntled about is that as a result of the Web build-out, most of the data is no longer in the database. You have this massive proliferation of XML and HTML, where data is just stored in the file. We need to build more of a file system approach to managing that data.

CRN : What other opportunities are on BEA's horizon?

SD : One of the places the industry has to step up is a better browser model for wireless clients. It has to support disconnected operations, higher latency and less reliable connectivity to those devices. There is very compelling technology that is going through the meat grinder right now to do next-generation Web browsing on mobile devices. We have a bunch of key alliances to take this technology forward. We're starting to see similar noises from other vendors. The basic idea is to marry asynchronous messaging with mobile devices running browsers.

CRN : So when all is said and done, why should people go with BEA?

SD : They should go with BEA because we're a more credible server and integration provider. Microsoft's .Net is a Windows-centric strategy, and Windows .Net is now the largest operating system and most opaque in terms of understanding it and trouble shooting it. The trouble the enterprise is feeling is how do they troubleshoot a Windows system online like they can do for Solaris or MVS. When Windows gets into a strange situation, you reboot it. That's not a great option when you have $2 million transactions pending on the server. In a nutshell, we're as business-critical as IBM but easier. And we're as easy as Microsoft, but much more business-critical.