A picture is worth 1,000 words. This simple truth is behind the success of various visual modeling tools that help business owners and IT experts reach a common understanding of needs. Last week, Kalido introduced the visual modeling paradigm to a new realm with the Kalido Business Information Modeler, which it bills as "the first and only solution that transforms the design, development, deployment and maintenance of BI infrastructures."

A component of the Kalido Information Engine, a data warehousing and BI infrastructure management environment, the Business Information Modeler breaks with the slow and cumbersome practices of static data modeling and custom programming that go into data warehouse development and ongoing maintenance.

"Data models are fairly complex, with lots of tables, joins, structures and funny label names, and they're not something you would show to a business person," says CEO Bill Hewitt. "Our Business Model approach provides a way to have a discussion with a business leader in terms that they understand."

BI analysts using the Business Information Modeler can sit down with the head of sales or another business professional to talk about how they want to see their information. The tool's "gesture-based" modeling environments lets the analysts quickly draw transaction systems and the data environment, with mouse actions quickly establishing the flow of information.

"If you drag the cursor from left to right, you create a dimension," Hewitt explains. "If you drag it from right to left, you create a transaction system. If you click the right mouse key and drag it from a transaction to a dimension, you create an association."

Once the general outline of the business model is completed, the BI analyst uses the modeler's control panel view to flesh out technical details, check errors and ensure that the model is valid. There's also an interface for DBAs and architects geared to bulk editing. The tool is is aimed at slashing the time needed to get to a live production data warehouse while also easing ongoing changes to data structured and downstream reports.

"The modeling tool will serve as a three-way bridge between the business owner, the data architect and the back-end database," says Xuhui Quan, a senior systems analyst at Abbott, a global pharmaceutical company and Kalido customer that is beta testing the modeling tool.

Quan says the Abbott's current practice is to draw out requirements in a Visio-like tool so architects can talk about requirements with business owners, "but then another analyst or IT person needs to reference the diagrams and develop the objects one by one. The modeling tool has a function that lets you turn the drawing into all the data objects needed to build the data warehouse."

The integrated modeling environment will cut Abbott's typical one-week development time in half, savings that will accrue every time there's a change in the warehouse, says Quan.

Kalido's complete Information Engine environment includes the core Kalido Dyanmic Information Warehouse, the Universal Information Director and Kalido Master Data Management, all of which are designed to work with existing databases, data warehouses and transaction systems, easing metadata management and ongoing maintenance. For example, the Universal Information Director integrates with Business Objects, Cognos and more than 30 other BI tools, automatically creating and maintaining Business Objects Universes, Cognos models and other metadata models and reporting infrastructures. "The Kalido Information Engine is interesting because of its ability to generate metadata for the major BI tools and in support of the ETL vendors," says analyst Henry Morris of IDC. "With three major BI companies being acquired in 2007 [Cognos by IBM, Business Objects by SAP and Hyperion by Oracle], the ability to provide this capability not just within a platform vendor's stack but across heterogeneous toolsets and databases is valuable."

Kalido's new Business Information Modeler is expected to be released in March. Other upgrades announced last week include extended support for very large dimensions, enhanced data staging management, with wizard-generated "feeds" that define how staging tables are loaded, and a new reference data loader, which works in conjunction with the underlying database server to improve processing performance. In addition, a new suspense processing capability automatically routes non-conforming data to suspense tables where errors can be identified and corrected.