A Virtual Point of ViewDespite limitations, virtual approaches to data integration hold great potential for certain applicationsContinued from Page 1 In my view, EII is appropriate to queries that require extremely fresh data, when the business end user can afford to wait a few minutes — even an hour — to get the data. For instance, many corporations run operational reports overnight, so executives can see yesterday's corporate performance. As an alternative, a few companies run operational reports via an EII platform, so executives can examine corporate performance every two or three hours. Some of these companies replaced an operational data store (ODS) in favor of EII. Venues and VendorsAside from using EII as ODS replacement, the virtual views of an EII platform can simplify or enhance query operations in a variety of other practical venues. For instance, a reporting view might facilitate reports that hit multiple, heterogeneous data sources. Or a customer view might provide up-to-the-moment information for customer-facing business processes, such as call center or sales. Furthermore, a supply chain can be a vast process, seen only in pieces, whereas a supply chain view may provide visibility across it and into it. And finally, a list of outstanding orders is a common information request, so a view might pull together this rapidly changing data for the purpose of self-service (perhaps through a portal) by employees, customers, and partners. EII is served by a vigorous vendor community, which includes pure-play EII vendors (for example, Enosys Software Inc., Ipedo Inc., Nimble Technology Inc., MetaMatrix, and XAware Inc.), as well as midsize vendors with enterprise application integration and EII products, such as BEA System Inc. and Vitria Technology Inc. IBM is, at the moment, the vendor with the most complete product and strategy roadmap for EII. IBM DB2 Information Integrator includes best-of-breed components for federated database management (DB2) and real-time integration (WebSphere MQ). It has mature capabilities for distributed query optimization and native connectivity to back-end databases that smaller vendors can't easily reinvent. Unlike other vendors, IBM's concept of EII encompasses both structured and unstructured data, as well as bidirectional, heterogeneous database replication. All these capabilities raise the bar on what an EII platform must do, giving IBM a lead that other vendors will find hard to close. With so many small vendors in this new IT product category, market consolidation is inevitable in the next two years, as smaller vendors get acquired or go out of business. Even so, the numerous releases coming from vendors this year are advancing the capabilities of EII, making 2003 a good year for evaluation and possible implementation. Virtues and VicissitudesEII presents potential benefits. It can accommodate unanticipated queries that need odd combinations of source data. Some queries run so infrequently that they don't merit a costly data integration process for precaching data. Once data is moved, it becomes stale, whereas queries against a virtual view fetch fresh data from sources in near real time. The virtual world of EII is usually faster and cheaper to design, implement, and maintain than the physical world of data-movement-intense approaches to data integration. However, EII also has its share of challenges. Queries into a virtual view take longer to execute than those directly against a physical database. A change of data structure in a source or target system may break the views and other metadata constructs in the EII platform. Queries through an EII platform put loads on source systems at unpredictable moments, thus complicating load balancing and capacity planning. For moving and integrating massive data sets, ETL has no substitute. Yet, for relatively small amounts of time-sensitive information that demand real-time data integration — as is often the case with data about customers, fulfillment, and corporate performance — EII's virtual point of view is an appropriate and complementary solution. Philip Russom, Ph.D. [www.PhilipRussom.com] is a Giga Research Director at Forrester Research Inc., where he provides advice to user organizations about business intelligence, data warehousing, and data integration. RESOURCESBEA Systems Inc.: www.bea.com Enosys Software Inc.: www.enosyssoftware.com IBM: www.ibm.com Ipedo Inc.: www.ipedo.com MetaMatrix: www.metamatrix.com Nimble Technology Inc.: www.nimble.com Vitria Technology Inc.: www.vitria.com XAware Inc.: www.xaware.com Related Article at IntelligentEnterprise.com: "Beneath the Waterline," May 7, 2001 "Orchestrating a Torrent," Jan. 8, 2002 "Fish Eat Fish — Another ETL Vendor is Swallowed," July 11, 2002
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| ||||||||||||||||||||||||||||||||
