The Sum of Data's Parts

Dump the pieces of several different jigsaw puzzles, a canister of Lincoln Logs and the contents of an Erector Set on the floor. Throw away a handful of the pieces for good measure. Now create something that's beautiful.

If the job seems daunting, you're feeling the pain of the solution providers tasked daily to integrate loads of disparate data, all the while wrestling with different interfaces, varying levels of data quality and gaps in compatibility.

Looking to solve some of these headaches for solution providers that seek to play in the growing but challenging data-integration space, vendors are prepping a new generation of server applications that promise to simplify the way disparate business data is tied together.

With the services-oriented architecture (SOA) now defining modern-day software and applications, it's only fitting that the SOA approach be applied to data-integration tools. The result, vendors and their partners say, will be new and better ways to integrate data by allowing more rapid development and simplified connectivity through a consistent user interface.

Ultimately, the mission for partners is simple: help customers provide data faster from more sources, and with better data quality. That means tying together data-cleansing tools and disparate data repositories.

IBM, Oracle, and Sybase, among others, are in a race to bring out the first SOA-based platforms that provide data integration and tools to improve data quality. IBM this month is taking the first step with the release of its IBM Information Server, a tool it describes as a "catalyst" for its Information On Demand initiative. But this is no dip-your-toe-in-the-water, entry-level kind of software infrastructure play. Getting started with IBM Information Server (IIS) will run customers close to $100,000.

For its part, Oracle is readying its data-integration debut. Meanwhile, the company bolstered its data-integration efforts last month with the acquisition of Sunopsis, a leading provider of heterogeneous data-integration tools.

Also imminently due out is version 1.0 of the Sybase Data Integration Suite.

Other software vendors going down similar integration paths include Business Objects, Informatica, NCR/Teradata, Tibco and SAS. Still, IIS represents the biggest advance in tying together such functions as data quality, ETL (extract, transform, load), enterprise integration and data quality under a tightly integrated SOA and metadata infrastructure, says James Kobielus, an analyst with Current Analysis.

"IBM has the first fully integrated SOA-based, data-integration suite on the market," Kobielus says. "This suite consolidates their value proposition into a unified platform offering that they can license as a single integrated solution to support a vast range of enterprise integration requirements."

IBM launched the Information Server at its first-ever Information On Demand conference in Anaheim, Calif., last month. The software is the confluence of years of internal development and some key acquisitions, notably Ascential. It's due to ship this month. Ambuj Goyal, general manager of information management for the IBM Software Group, says the tool ties together previously disparate functions.

"Before IIS, you would have to make a specific call--JMS, ODBC or JDBC--to another source of information," Goyal said at the launch event. "With IIS, you can do a single call. You can get info from multiple sources. We hadn't solved the information problem for SOA projects. We had solved the process problem when we announced the WebSphere process server last year. Now we have a common server to solve the information problem."

The fact that IBM has tightly integrated IIS with the WebSphere Registry and Repository (based on the UDDI standard) will simplify the sharing of metadata and other components of IBM's growing portfolio, Kobielus says. He also notes that the new software publishes all data-integration functions as Web services, thereby allowing customers to leverage IIS into more complex data-integration and master data-management repositories.

"They've done much like they did with application servers," says Rupert Bonham-Carter, an associate vice president at BI vendor Cognos, who oversees his company's alliance with IBM. "They hardened the category for information servers. They brought to bear existing parts, but they rearchitected them in a common way using a modern service- oriented architecture, to take what were disparate pieces and provide a single offering for a broader category."

NEXT: So, who needs all this? The short list includes the growing number of businesses looking to create a single view of customers, those looking to simplify business intelligence and organizations that require accurate, real-time, up-to-date information for compliance or other mission-critical functions.

Big Blue says that IIS unifies five key areas. It provides a single user interface across all functions segmented for analysis, development and administrative functions; shared services that centralize core tasks such as security, user administration and reporting across the various product lines or repositories--these common services include metadata services which provide unified service-oriented access and analysis of metadata; common metadata, shared definitions and syntax across business and technical domains, thereby reducing project delivery times and improving overall confidence in information; unified connectivity; and a common parallel processing engine, which IBM says eliminates processing capacity as a barrier to achieving project results.

For VARs such as Lincoln Park, N.J.-based Genware Computer Systems, such improvements in data integration should mean faster delivery of solutions.

"We should see a faster time to access information that's out there," says Sherlock Holmes, Genware's president. In terms of building data warehouses and other BI applications, these data-integration platforms will eliminate the need for ETL and EAM tools, among other components.

"This is a confluence of many products," Holmes says. "All of my accessibility to information can now be done through this one central service, which will feed my BI front end and also feed other applications that rely on this data."

Holmes and his key BI partner Cognos say they will also support Oracle's data-integration offering when it comes out.

At least for now, Holmes and others say IBM has an early lead, though most agree that it will be short-lived. Nonetheless, Holmes says he expects to see significant demand. "The SOA initiative and the ability to bring [disparate repositories] together into a single environment will be a crux for speeding up delivery, and that's a big thing for us," he says. "I think we're going to see a huge part of our business driven by this type of architecture."