In his latest book, The Data Webhouse Toolkit, Ralph Kimball notes that the Web revolution has "propelled the data warehouse out onto the main stage". Data warehousing and e-business are converging, with the data warehouse taking on two personalities. In one sense, the Web provides an ever-increasing and unprecedented source of valuable behavioral data about prospects, customers, business partners and employees.

At the same time, warehouse architecture is changing from client/server to a fully Web-centric environment. The reasons are obvious. The Web is global, available 24 hours a day, all business constituents are connected and the only prerequisite software is a standard Web browser. A Web-centric architecture must deliver all of the functionality data warehouse users have traditionally expected, including simple query and reporting, data mining and OLAP analytic applications, all within the ubiquitous Web browsers.

By linking clickstream data with information collected from brick-and-mortar operations, organizations are becoming more customer focused and developing true enterprise profitability models. In addition to understanding the profitability of customer segments, products, channels, operating entities and promotions, businesses are gaining insights into the activities and behaviors that actually drive profitability. Indeed, the Wall Street correction that occurred last spring has underscored the need for businesses to understand customer profitability and demonstrate ROI of their online initiatives. It is no coincidence that International Data Corporation is projecting the data warehousing market to exceed $11 billion by 2003.

Online analytical processing (OLAP) is widely recognized as the essential component of a data warehousing architecture for enabling business analysis applications. Specifically designed to perform sophisticated calculations and comparisons across many dimensions, OLAP servers provide the instant response times users need to interactively explore dynamically changing business conditions. As Kimball notes: "OLAP represents the next evolution in decision support systems because OLAP has far more powerful capabilities than SQL."1

As business analysis continues to become a more strategic enterprise initiative, application developers are demanding more from OLAP technology. Applications (such as supply chain analysis; workforce planning and variable compensation analysis; marketing, sales and services analysis) must be deployed over intranets and extranets and may be accessed by thousands of users. Like all software components of a fully Web-enabled architecture, an OLAP platform must be highly scalable, cross platform, standards compliant and have broad third-party support.

In the converging world of e-business and data warehousing, the importance of meta data interoperability expands exponentially because of the decentralized and stratified nature of Web architectures. The lack of meta data interoperability between best-of-breed data warehousing tools, repositories and databases has been a long-standing and vexing challenge for warehouse designers, administrators and users, even in the pre-Web era. The way meta data is stored, structured and defined is unique to each vendor's specific tool or application. This makes it difficult to deploy and administer systems. Data warehouse administrators, for example, need to understand how changes to job schedules impact the availability of specific reports and which users will be affected. Developers need to know how a change to a source system table impacts extraction, transformation and load processes. End users rely on meta data to navigate and understand the data they are working with.

Application developers have historically worked around the problem of meta data incompatibility by creating custom bridges or relying on vendors to develop point integration solutions between specific tools and applications. For developers, these approaches delay deployment and increase the cost of maintenance and administration. For vendors, the effort associated with developing and maintaining meta data bridges represents a diversion of programming and QA resources that could be channeled toward developing value- added functionality. Furthermore, there is an ongoing need to rework and retest these bridges each time either party releases a new version of their product.

Although vendors have long recognized the need for a meta data standard, efforts to agree and standardize on a common approach have not borne fruit until recently. The rapid acceptance of XML as the universal language of e-business has been a key factor in alleviating some of the barriers to a common meta data standard. The Object Management Group's (OMG) ratification of the Common Warehouse Metamodel (CWM) represents a major advancement toward an approach that uses an XML interchange format to simplify the process of designing, building, administering and navigating data warehouse infrastructures. CWM is based on the OMG's Meta Object Facility (MOF), a standard that defines an extensible framework for creating and accessing meta data models.

The industry adoption of CWM will be aided by the proven success of OMG and its 800 members in producing and maintaining specifications for open systems since 1989, as well as by the broad industry support that CWM has already garnered. The CWM standard was co-submitted by Hyperion Solutions, IBM, NCR, Oracle and Unisys with guidance, review and ratification from OMG's extensive industry and vendor membership.

Another important complement to XML and CWM is the establishment of Java as the defacto programming environment for Web-based applications. The Java 2 Platform, Enterprise Edition (J2EE) offers data warehousing and application developers a scalable, platform- independent programming language that is managed and advanced through the Java Community Process 2.0 (JCP) program, a community-based process for developing Java technology specifications and their associated reference implementations and compatibility test suites.

The J2EE platform will soon be enhanced by a pure Java OLAP API (JOLAP) that will support creation, storage, query access and maintenance of data and meta data in OLAP servers and multidimensional databases in a vendor-independent, cross-platform manner. The interdependency of e-business and analytic applications heightens the urgency and significance of providing the Java development community with a simple, pervasive mechanism to work with OLAP data structures. As of June, there have already been close to three million downloads of Java 2. For this rapidly growing community of Java developers, JOLAP brings to OLAP databases what JDBC provides for relational databases: cross-platform rapid application creation, standardized integration and easier maintenance. Initiated through the JCP by Hyperion Solutions, JOLAP has been endorsed by IBM, Oracle, Unisys, SAS, Sun and Nokia.

Just a few years ago, data warehousing was an ancillary business activity intended to extend data in ERP, CRM, legacy and other operational systems with a single version of the truth for internal decision support. The Web has redefined the concept of the data warehouse, from both a functional and architectural perspective. The data warehouse has evolved beyond a simple query and reporting environment, to a Web-centric infrastructure for delivering mission-critical analytic applications that are vital for understanding and profitably enhancing the ability of businesses to acquire, keep and grow customers. Meta data is the glue that binds the data warehouse, OLAP is the engine that powers business analysis applications and J2EE is the development platform of the Web. The rapid and broad industry adoption of XML, CWM, J2EE and JOLAP are fostering the essential open systems that accelerate the delivery of e-business intelligence.

References
1 Kimball, Ralph and Merz, Richard. The Data Webhouse Toolkit. John Wiley & Sons. 2000. P. 265.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access