Choice Of Database Management System Has Implications For Years To Come, Idc Warns

EDP Weekly's IT Monitor, April 12, 1999

Choosing a database management system (DBMS) for the enterprise involves much more than making a tactical decision to solve an immediate business need. The effects of that choice will be felt within the organization for years to come, warns IT market watcher International Data Corp.

"The wrong DBMS choice can lock the user into a technology that does not serve the enterprise well and is difficult and expensive to change," says Carl Olofson, director with IDC's Database Management Systems research program. "The right decision can considerably alleviate the burden of maintaining and extending the enterprise information infrastructure."

According to a new IDC report entitled Database Strategies for the Enterprise: One-Stop Shopping Versus Best of Breed, the choice of database depends on many factors including indexing, recoverability, robustness, and scalability. Much of the decision will depend on what the organization demands from its database.

Large online transaction processing (OLTP) systems present different issues than do small ones, and decision-support databases are different from OLTP databases in planning, deployment, and use. In addition, large decision-support databases represent somewhat different challenges than do smaller, single-subject data marts. Some DBMS vendors say they can offer a single point of contact for providing database functionality in all areas. Others specialize in particular types of databases. Organizations need to decide if they want to use the single-vendor approach for simplicity or the best-of-breed approach to ensure they have just the right technology for each job.

"If one database is the key one and others need to be there but do not have equally exacting requirements, it may be reasonable to use the critical one as the basis for selecting a DBMS vendor to standardize on," Olofson says. "If, on the other hand, you have a variety of databases with different characteristics and special requirements, as well as an organizational need for each to fully satisfy those requirements, it makes sense to take a best-of-breed approach and choose a DBMS for each database on its own merits."

Olofson offers this final advice: "Building a database is never as easy as the vendor says it is. Make a short list. Check out the documentation. Try the product before you buy it; use as realistic a test case as you can within the resource and time constraints of the trial. In the end, go with a vendor you trust. It's not just a product; it's a partnership."