When managing complex database environments, IT vendors and buyers agree on the three top priorities: integration,...
integration and integration.
The challenge of integrating data is not a new one. However, restrictions on IT spending and staffing in the last year have placed even greater importance on knowing how to integrate existing systems rather than investing in new technology.
"Quite honestly, the complexity of how you choose to build applications, and systems, has mushroomed," said Stephen Hendrick, a vice president at Framingham, Mass.-based International Data Corp. "We've ended up with a massive number of islands of technology."
That explains why so many vendors seem to be bragging about their integration capabilities these days.
Seeing leading integration vendors such as SAS, Informatica and Ascential Software squabble over what several research groups estimate is a $1 billion integration market, major database management system (DBMS) vendors have recently announced integration initiatives of their own.
In early February, IBM Corp. introduced DB2 Information Integrator as a way to accommodate integration of structured and unstructured data from disparate sources.
Oracle Corp. has undertaken an enormous development effort to build integration capabilities into its 9i Application Server. Meanwhile, Microsoft's BizTalk extends the reach of SQL Server 2000 into many non-SQL data sources.
Still, these solutions are often tightly coupled to the vendor's database, and there are a slew of products that support an entire integration plan.
Hendrick pointed out that products ranging from extraction, transaction and loading (ETL) to data analysis and re-engineering tools all provide aspects of data integration.
For example, when Gartner Inc. examines data integrators, it considers ETL providers. SAS, Informatica and Ascential are obvious market leaders in this space. Even those vendors, though, can be categorized in various ways, as they compete with vendors such as Business Objects and Cognos, companies with strong identities in the business intelligence (BI) marketplace. Each of those five vendors holds a spot in Gartner's January 2003 ETL "magic quadrant," which shows current leaders and challengers in the data integration space.
The Gartner report states that BI players and major DBMS vendors are making life tougher for market leaders such as Ascential as they acquire and build ETL tools.
Separating EII from EAI
One way to simplify the world of data integration is to differentiate between data and application integration.
"When it comes to getting your feet wet in the whole issue of integration, there is a pretty strong belief that the path you probably want to take first is the data integration path," Hendrick said.
That means separating enterprise application integration (EAI) from enterprise information integration (EII), which is the taxonomy IBM assigned to its new DB2 Information Integrator.
Choosing a vendor/product
SearchDatabase.com site expert Pat Phelan, a data modeler for a large U.S. business services company, said he considers only those products which address EII as ones that are performing pure data integration.
Vendors such as Ascential fall into a different category, Phelan said.
"In terms of doing integration with an off-the-shelf, large-scale manageable package, I would consider what Ascential is doing to be a strong viable precursor to what the large database vendors are offering, but I wouldn't consider them to truly be a competitor," Phelan said.
Phelan said that a good starting point for IT managers considering data solutions is to take into consideration the type of data they are generating and whether they are already heavily invested in a single vendor.
"If you have a huge investment in Oracle, that increases the likelihood that the Oracle EII approach will be well-suited to your needs," Phelan said. "The same is true for Microsoft and IBM."
Philip Russom, research director for data integration at Cambridge, Mass.-based Giga Information Group, agreed that IT customers are often confused when faced with data integration choices. "A lot of products share attributes and overlap," said Russom. "There's ETL, EII, and the data replication folks -- Quest's SharePlex, Data Mirror. That's all data integration."
Russom credited IBM with having the most "open" approach to data integration among the top three database vendors, and he said that even customers running predominantly Oracle shops could benefit from the new DB2 Information Integrator -- depending on which sort of data they generated.
"Going with the niche vendors means using ETL," Russom said. "That process can take weeks. EII, like IBM is using, collects data and integrates it within a few minutes."
But, Russom said, there is a trade-off. "If you want to move huge amounts of data, it has to be done with an ETL, too," he said. "For relatively small amounts of data, you want an EII solution."
FOR MORE INFORMATION: