In The Gorilla Game (255), Moore distinguishes five stages in the evolution of computing:
1. mainframe
2. mini computer
3. PC/desktop
4. client-server
5. Web-based distributed computing.
Moore predicts that Wave 5, which we are approaching and for which BEA aims to be the enabling software, is likely to produce a supercategory of software called enterprise client/server applications (GG, 257). This infrastructure supports the new world of Online Transaction Processing (OLTP) and will, in Moore’s opinion, be the biggest driver of investment in computer systems in Wave 5.
At issue is the role BEA will play in Wave 5. It is likely to emerge either a gorilla or a strong chimp. This is not a hospitable place for a king to build a castle.
Value Proposition
WebLogic 6.1, using J2EE and incorporating Tuxedo and EJB, becomes the dominant enterprise enabling software and platform for Wave 5. With WebLogic Collaborate and WebLogicIntegrate, BEA supplies the building blocks or architecture for this “fifth wave.”
When Shapiro and Varian published Information Rules: A Strategic Guide to the Network Economy in 1999, CEO Coleman had his entire management read it. The reason is pretty clear. Networks require that you identify your natural allies early in the game and that you cultivate their cooperation. WebLogic’s open compatible communication and interoperability standardization create value across a range of OLTP and on a global scale. Independent software vendors (ISVs) and enterprises are drawn to WebLogic because other ISVs and enterprises have been drawn to WebLogic.
Technology/application or enabling?
BEA has feet in both the applications and the enabling ponds. While Enterprise Java Bean features personalized applications components, BEA has increasingly moved to solidify WebLogic’s position as the platform for the distributed computing model in the e-commerce realm.
Competitive advantage
BEA’s list of ISV’s is heads and shoulders over its closest competitor (IBM), and it has at least twice as many customers. BEA does not try to lock customers into a database (Oracle), a mainframe (IBM), or a closed operating system (MSFT).
A major competitive advantage is created by the “enhancements” integrated into WebLogic that customize the platform according to a customer’s specific e-commerce needs. Oracle, NCR, and Salesforce.com, for example, have offered “out-of-the-box” CRM solutions. But as YOUcentric CEO Don DiLoach, a BEA partner, points out: “To leverage top line initiatives, you need to get people to buy from you as opposed to buying from your competitors.” Out-of-the box relationship management systems take the unique dimensions of a business and subject them to the overall structure of an imposed system. Bank of America, Federal Express, and Seagate Technology are joint YOUcentric and BEA customers running YOUrelate on WebLogic 6.1.
biz.yahoo.com
The flexibility BEA offers its customers and ISVs—the number of components and open architecture—makes it easy to adopt and less expensive to write software off the BEA platform.
Result: Huge network effects are tripped and switching costs created. For this reason alone, IBM and Oracle are not likely to overcome BEA’s lead. Most of this lead stems from BEA’s value chain and the network effects it triggers. But not all. In benchmark tests, BEA’s WebLogic Server 6.1 proved to be up to four times as fast as IBM’s WebSphere Application Server 4.0 and executed up to 14 times as many transactions per second on key individual JavaBean tests. Compared to Oracle’s new 9iAS applications server, which does not offer EJB clustering, BEA’s WebLogic Server is 54% faster when measured against Oracle’s own benchmark (Pet Store).
bea.com
bea.com
Not only do Oracle and IBM face the Innovator’s Dilemma, they are competing against BEA’s first mover advantage. Soon after BEA licensed Java, it acquired WebLogic (September 1998). The next month BEA announced a development program for ISVs. By December, BEA had won two Best-Web applications awards from Java publications for its new WebLogic application server. BEA won these awards in part because it was the first and only server to comply fully with Java specifications.
MSFT’s .Net is the big question. WebLogic, in contrast to .Net, is proprietary, but open and operates using J2EE. This is the future platform battlefield. With MSFT’s huge installed base, it is primed with .Net to test the first-mover advantage BEA enjoys. The ensuing battle will determine whether the architecture of Wave 5 remains open.
Steve Garone, IDC vp of app development and deployment research, comments that he is seeing some platform alignment in the applications server market with IBM customers going with WebSphere, etc. BEA, being platform unaligned, could be at risk although the trend is early and may be skewed by the way that IDC looks at revenues. IBM’s business model is based on service, and it licenses WebSphere for free. On the other hand, the industry is standardizing on Java, which may be a move to neutralize MSFT’s attempt to extend its gorillahood into the enterprise app server market.
Discontinuous Innovation: WebLogic E-Business Platform and the transformation of “middleware” to an enabling software.
BEA’s discontinuous innovation comes in taking middleware to a higher level, what Moore calls in LOTFL a “platform product”—enabling software that acts as an operating system for a distributed computing environment. Moore predicts in GG (p. 74) that middleware is “likely to evolve in a manner akin to the deeper enabling software layer, privileging a single, dominant gorilla blessed with an ever-increasing market share.”
In client-server architecture, the number of connectors needed can be calculated by the formula N2 –N where N=the number of applications. This is an expensive, clumsy, and inefficient mode as systems grow in complexity. With 50 apps, for example, 2,450 connectors are needed. [Thompkins, Platform Wars]
Discontinuity: The WebLogic platform provides an integration solution, based on XML messaging, that permits all applications to communicate in the same language, coordinates transactions within and across B2B partnerships, and sends events across apps automatically to update interested parties in real-time. Compared to the point-to-point client-server computing model, the n-tier configuration produces an order of magnitude improvement in performance, efficiency, interoperability, integration—and, last but not least, cost.
Proprietary Open Architecture: WebLogic Server 6.1, introduced in June, integrates Application, Transaction, Integration, Process and Workflow, Portal, and Web Services Servers. WebLogic 6.1’s n-tier configuration meets e-business imperatives for a comprehensive platform by decoupling the transparent business layer (which contains vertical applications) from the behind-the-scenes heavy lifting that takes place in the back-end layer’s horizontal middleware (which optimizes net-centric functions). The architecture has three basic layers:
presentation layer: allows user to interact with the system
business layer: the brain that governs data logic and generic/specialized functions (real-time deployment, performance optimization, etc.)
administration layer: graphical interface that deploys, configures, and fine tunes applications accessible by a browser
As an enabling software, the architecture of WLS 6.1 simplifies and standardizes the development, deployment, and integration of applications by eliminating overnight batch downloading, point-to-point integration, and coding of QOS service functions. Based on open standards, these in turn drive increases in productivity, portability, reusability, and interoperability without creating proprietary lock-in for vendors, developers, or customers.
The BEA Business Process Manager (BPM) embedded in WLS 6.1 enables a business analyst to drag-and-drop flowchart elements of business processes without recoding software. According to Giga, “Companies can expect to accelerate their development/integration time by 50% to 150% using workflow instead of coding business logic into applications.” With BPM’s process engine, the automated workflow process is executed and managed dynamically, updating asynchronous interactions whether the business transactions flow over minutes or days. Thus, human skills are leveraged from project to project, and the reusable components can be selected or integrated to form new solutions.
BEA’s whole strategy is built on the concept of proprietary but open architecture. Tuxedo’s success demonstrates an early deployment of this strategy. When BEA acquired Tuxedo, IBM had its own OLTP, called the Customer Information Control System (CICS), which handled customer transactions on mainframes. According to PaulPhilip on the Fool, BEA positioned Tuxedo to surround and interoperate with CICS. Tuxedo not only communicates with CISC, but also cooperatively interacts with it to yield a single unified transaction. Because in large part of the multiple benefits of this interoperability, Tuxedo became the market share leader for OLTP on both Unix and mainframes.
The second example is the intersection of Java and WebLogic as BEA’s infrastructure solution evolved. After fully implementing all Java specifications, Coleman and BEA management made the decision to let the standardization of new Java protocols drive and extend the standardization of applications on WebLogic.
This strategy contrasts with those of BEA’s competition. Oracle and IBM have opted to retain their legacy proprietary approaches to locking in customers: Oracle through its database management and IBM through its mainframes. IBM has been competing with BEA in the space for a while. Oracle’s decision to play in this sandbox is questionable at best. Not only is there an entrenched player with dominant share, a strong value chain, and acknowledged expertise (BEA) and another entrant (MSFT) with prodigious assets, Oracle will be directly competing with its customers, a move that might encourage some to jump ship for database to IBM or Microsoft.
MSFT is obviously positioned differently for the upcoming battle than IBM, ORCL, or SUNW. While MSFT protects its Windows environment in .Net, the magnitude of its installed base should prevent it from suffering the Innovator’s Dilemma to the same extent as Oracle and IBM in the space.
The fight for control of the platform for Web-based enterprise distributed computing will, in our judgment, pit BEA, the anti-lock-in potential gorilla, against Microsoft, the lock-in specialist and gorilla of Wave 4. MSFT’s recent announcement of offering greater flexibility to OEMs for Windows (and XP), while providing options to add/remove icons and access to Internet Explorer, is, to our mind, an overture to DOJ/Appeals Court [and a pr/legal positioning ploy prior to XP’s release]. It is not a signal of any fundamental platform decision, certainly not one that will greatly impact .Net.
biz.yahoo.com
While BEA does not have the cash horde of MSFT, it continues to strengthen its technology through in-house development and acquisition. It spent $89.2m, $61.0m, and $42.1m in fiscal ’01, ’00, and ’99, respectively, on R&D—quite apart from the R&D it gained through acquisition during that period.
Switching Costs: Switching to any e-business infrastructure is hugely expensive for enterprises. Switching from a mission-critical network-centric platform can be catastrophically costly. Selecting WebLogic as an e-business app server is part of a lengthy and mission critical process that takes a company from a client-server environment to a distributed computing model. BEA’s solution has some compelling attractions—its openness and technological superiority in several key areas. Once an infrastructure is installed, users are hesitant to switch. And as more enterprises make that choice, switching costs escalate.
Barriers to Entry: It is relatively easy for an independent software developer to write J2EE applications. It has been suggested that this fact shows that BEA has created few barriers to entry. On the contrary.
Developing applications is not BEA’s primary focus. It facilitates applications on the WebLogic platform—whether developed by BEA or ISVs. The ease with which these applications can be produced off the platform should not be confused with the difficulty and expense of producing the platform—especially one with all the proprietary bells and whistles or interoperability that BEA has built into WebLogic.
The value chain itself creates huge barriers to entry here, and as more and more ISVs and companies like SEBL or PeopleSoft standardize on WebLogic, the barriers to entry for a potential competitor get higher and higher. The recently announced BEA deal with Dell to preinstall WebLogic on many of its servers, for example, creates another opportunity with obvious network benefits. Similarly KPMG Consulting is building integration solutions for the high-tech industry with its Configure to FIT framework using BEA WebLogic Integration to fasttrack successful implementations for its clients. Message 16052127
Any new entrant into the app server market will find it difficult terrain. Even Oracle’s entrance here is dubious. This is not a mature market that can be attacked “from below” or at the margins. BEA enjoys first-mover advantage among the established Java companies. And any new entrant would not only have to contend with the Java contingent, but sidestep MSFT’s .Net initiative.
Value Chain: BEA participates in a host of partnerships, SEBL being the most recent and most strategic. Nokia (NOK), Documentum (DCTM), PeopleSoft (PSFT), and Sterling Software (SSW) are among others. Through BEA’s Star Partner Program, system integrators (SIs), application software providers (ASPs), and ISVs partner to accelerate time to market, reduce deployment costs, shorten sales cycles, and increase revenues. To date 1,900 partners participate in the program, with additions at the rate of about 100/month. Channel partners currently account for about 25% of BEA’s sales, with a goal of 40% by 2003 and an ultimate goal of 60% or more to come from indirect sales.
Total number of licensees of BEA products and solutions is greater than 9,400 worldwide (10K). Over 180,000 ISVs are registered at Web site BEA maintains for them. They downloaded over 900,000 copies of BEA software in fiscal 2001. Thirty-eight thousand WebLogic programs were downloaded at the company Website for a free trial in Q3’01 versus 12,000 in Q4’98 (10K).
Networks become “bigger” by increasing their reach and “better” by increasing their richness, which may generate even more explosive value than reach. BEA increased network reach in terms of the number of business applications that could be written off its platform by abstracting the QOS functions that were hardest to code into the platform. It increased richness by integrating its Business Process Model and more servers into the platform.
BEA’s strategy exerts both direct and indirect network effects within the value chain. Software designers design applications on top of WebLogic because it’s the most popular platform; then customers standardize on WebLogic because that’s where the greatest number of applications are available. Using the dynamic feedback of BPM’s drag-and-drop modeling, business analysts and software developers are building a future out of Java building blocks. Simple, but accurate business logic embedded in the software is reused and recombined to model ever more complex business processes.
Network effects of a standard platform and the numbers of applications that could be written on it drove Microsoft’s tornado in Wave 4. BEA’s management is looking to harness the same dynamic for Wave 5. The control BEA exercises over its value chain is not yet at gorilla strength, however. Schwab, for example, uses both WebLogic and IBM’s WebSphere.
Chasm/bowling alley/main street
With the acquisition of Tuxedo, BEA started to cross the chasm. Having enhanced the WebLogic Server (WLS) and using J2EE, BEA grafted Tuxedo, written in C++, onto the WLS platform. This became a springboard to knock over the pins in two visible and sizable bowling alleys: the telecom and financial markets.
Leveraging this success, BEA moved on to other bowling alleys. In quick order, it succeeded in capturing the overwhelming majority of Fortune Global 500 companies and across a broad market spectrum: telecommunications, financial services and banking, pharmaceuticals and health-care, computer/office equipment companies, insurance, aerospace, airlines, motor vehicles.
Hypergrowth/tornado?
Not yet. But giving off signals. Geoff Moore is a long-time advisor, and the company is clearly positioning itself to grab market share.
Growth rates for the product category of enterprise application servers are in tornado range. After increasing 110% in 1999 ($975m), revenues accelerated to 128% ($2.2 billion) in 2000 (IDC, July 2001). Giga forecasts that the market will reach $9 billion by 2003, while Forrester Research estimates that B2B electronic commerce will grow from $604 billion in 2000 to $6.3 trillion by 2004.
BEA has knocked over pins in successive bowling alleys: telecom, financial services, transportation reservation, etc. The real Gorilla game for BEA will be played out as it competes to become the standard platform, the enabling infrastructure, for e-commerce in Wave 5—similar to what Window became for Wave 4.
The possible tornado products were introduced in the last six months. In December 2000, BEA offered a complete reprogramming of WebLogic, still in J2EE, and called WebLogic Collaborate. Then in June 2001 came WebLogic 6.1. These products take on the architecture of Web-based e-commerce.
In the May 28, 2001 issue of VARBusiness, BEA Chairman and CEO Coleman said: “There’s no doubt that the infrastructure market is the biggest market there is in software….As far as software goes, the winner of this has the opportunity to be the largest software company in the world.”
This WILL be a tornado market. It remains to be seen whether it will be BEA’s.
Resources: Matt Thompkins’ “Platform Wars: The Next Generation” is a must read. It not only builds on G. Moore’s five stages in the evolution of computing, but positions BEA in the fifth stage: distributed computing. boards.fool.com
Foolish research:
Introductory post boards.fool.com
Compleat FAQ boards.fool.com |
