Intelligent Enterprise

Seeking to broaden its identity beyond being a niche player in data warehousing, Composite Software last month introduced release 4.5 of Composite Information Server (CIS). Since its founding five years ago, the company has been known for enterprise information integration (EII) technology, which enables organizations to access data from heterogeneous sources without the extraction, transformation and loading (ETL) steps typically necessary for data warehousing. (In some cases, CIS complements ETL.) Using query optimization, caching and metadata capabilities, CIS acts as a middle tier that combines data and presents the results to users as relational views or that supplies other visualizations and abstractions required by front-end applications and dashboards. The new release updates CIS to support service-oriented architecture (SOA) and emerging data services.

Data virtualization, the central thrust of CIS 4.5, has always been the key concept behind EII: that is, to give users access to data in traditional formats and other information without having to deal with the idiosyncrasies of each source. Through the introduction of services to access Teradata servers and legacy mainframe systems, as well as new technology called Composite Active Cluster (CAS), Composite is positioning CIS as the middle tier that enables virtualization much more broadly than just the data warehousing uses that EII has served. CAS can group multiple deployments of CIS into clusters to increase scale and for fail-over and disaster recovery; clusters can share a cache to hold result sets, which reduces the traffic between CIS instances and back-end data sources for frequent queries that, for example, produce parameterized reports.

Upgrades to Composite Studio, the company's graphical development environment, enable users to create data services, which can fit into the SOA to query sources for information about customers, orders and other business-oriented objects of interest. If the data services adhere to the plug-and-play goal that drives standards-based SOA implementations, organizations will be able to reuse the data services as components of different applications. This will save time and resources currently spent writing the same queries and data integration routines over and over again.

Virtualization is an important information management goal for organizations that have numerous sources from which they must integrate data easily and quickly. While the complexity of implementing SOA remains a challenge, Ventana Research believes that organizations should evaluate Composite's new release as a potential infrastructure for delivering information integration and data services that require more support than traditional data warehouses can supply.

David Stodder is vice president and research director - information management and IT performance management at Ventana Research . He was previously editor-in-chief and editorial director of Intelligent Enterprise.