IBM, Oracle Wrangle Over RDBMS Manageability Claims
Oracle’s claim that 9i is cheaper to manage than DB2 may not be all it’s cracked up to be
If a recent study from a Silicon Valley competitive intelligence firm is to be believed, Oracle Corp.’s 9i database is easier and less costly to administer than IBM Corp.’s DB2 platform with equivalent management workloads.
It could all depend on the meaning of the word “equivalent."
Late last month, Oracle touted the results of a new study that pitted 9i R2 against DB2. The study purported to determine how much time and effort database administrators (DBA) must expend in each environment to complete 12 basic data management tasks. The study, which was conducted by Rauch Associates Inc., a Silicon Valley-based competitive intelligence firm, appeared to validate Oracle’s claim that 9i R2 is cheaper and easier to manage than DB2.
For example, the Rauch study found that Oracle 9i R2 required 41 percent fewer administrative steps than DB2 8.1 to complete the same scenarios; that Oracle 9i R2 required 54 percent less time than DB2 to perform equivalent tasks; and that Oracle 9i R2 administrators could complete equivalent management workloads in 55 percent of the time required by DBAs in DB2 environments.
Moreover, the Rauch study put a price tag on Oracle 9i R2’s superior manageability: The increased efficiency of Oracle DBAs would result in cost savings of $37,054 per administrator in the first year of ownership. These cost savings would continue to accrue in subsequent years.
As it turns out, the Rauch study was sponsored by Oracle, which sponsored a similar study—also with Rauch Associates—in early 2002. Moreover, acknowledges Ken Rauch, a principal with Rauch Associates, his company didn’t actually come up with the idea of comparing Oracle 9i R2 and DB2 8.1 on its own. Nor, he concedes, did Rauch Associates determine the study’s methodology and management scenarios. Instead, Rauch says, “This is strictly a situation where the folks at Oracle said, ‘Hey, if we do these tests, we come out faster. Let’s get an independent company to look at what we did and document what we did.’ And that’s where we came in.”
It’s an important point, suggests Carl Olofson, program director for information management and data integration software research with International Data Corp. (IDC). “The methodology was basically to certify the procedure that Oracle provided them. We could assume that Oracle provided them with a scenario that would be favorable to Oracle.”
Jeff Jones, IBM's director of strategy for data management, echoes Olofson's point. “The main problem here is that the way an Oracle DBA with Oracle would approach a problem is many times vastly different than the way a DB2 DBA would approach the problem with DB2,” he notes. “So no attempt was made to see if the way one would approach the problem in an Oracle world is the same in the DB2 world.”
That’s not entirely true, Rauch protests. Oracle did propose the study, along with its methodology and management scenarios, he allows, but Rauch Associates also made an effort to ensure that the DB2 scenarios used in the study were efficient. “We didn’t just mindlessly say here’s how you do it in DB2 and hope for the best,” he stresses.
Instead, Rauch says, the DB2 scenarios—which were submitted by Oracle—were based on a list of best practices published by IBM itself. In addition, Rauch Associates tapped the expertise of several platform agnostic consultants—DBAs who support both Oracle and IBM databases—and IBM Gold Consultants to confirm the time estimates and test algorithms, Rauch says. “We can’t say that there aren’t more representative functions for DB2. We thought they looked reasonable, but we didn’t focus too much on that, because the document was very specific [about how to do these scenarios].”
Rauch Associates did not solicit a list of recommended steps directly from Big Blue, however. On the other hand, Rauch acknowledges, Oracle created and submitted its own list of steps for each scenario. “We didn’t really focus too much on the Oracle side, because we assumed that they would show themselves in the best possible light and we just wanted to ensure that what they were representing about IBM, at least in that context, appeared to be true.”
Rauch points out that Big Blue’s Tivoli management tools—which might otherwise ease the administration of DB2—aren’t shipped with the out-of-the-box database, and are available separately, at a price. As a result, he claims, DB2 DBAs must shift between GUI-based tools and command line tools to accomplish different management tasks. In Oracle 9i R2 environments, Rauch writes, Oracle’s Enterprise Manager can perform even the most complex tasks.
In addition, some of the test scenarios—such as the tuning of problem SQL statements—favor Oracle 9i R2 for the precise reason that it features a GUI tool designed for this purpose. DB2 has no such tool, Rauch writes, which means that problem SQL statements must be tuned by trial and error.
The upshot, observes Mike Schiff, a principal with data warehousing consultancy MAS Strategies, is that the conclusion of the study—that Oracle is less expensive to manage and maintain than DB2—should be taken with a grain of salt. “Vendors typically don’t publish studies that show that they came in last, so it’s probably important to keep in mind that your individual mileage may vary by a function of how well your DBA drives the database."
Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.