Worth Keeping is an occasional glance at technology that has been upgraded, replaced or phased out but still has a valuable place within the user community.
Eight is enough, or so goes the cry of many database administrators who are clinging to Oracle 8 or 8i and aren’t
upgrading to the new 9i software.
The earlier version has many features that still make it an attractive package, according to Peter Smith, president of the Ottawa Oracle User Group. These include partitioning, replication, enhanced memory management and locally managed table spaces. (Incidentally, some of these features are holdovers from version 7).
He also points out that there’s many third party software applications for 8 to choose from, like financial packages and query or monitoring tools. And he cautions some of these programs might not run on 9i.
“Just because something comes out for 8 today doesn’t mean you can install it on 9i tomorrow,” he says.
He adds that another trump card in 8i’s favour is that it’s now “very stable” and anticipated to run well for quite some time.
Joanne Pomalis, an independent consultant who helped implement Oracle 8i for the Bank of Canada’s data warehouse, says that the bank is sticking with the earlier version to ensure all of its applications are consistent with the database software.
“There were a number of features in 8i that we were (also) interested in,” she adds. “One was materialized views, the other was partitioning. I haven’t looked too deeply into 9i, but I don’t believe there’s anything in that release we’re dying to have.”
Oracle Corp. is trumpeting 9i’s increased scalability — the ability to change size or configuration — through clustering-technology as a main drawing card. But it appears there isn’t the demand for the extra space yet, which currently limits the feature’s marketability.
On top of that, version 8 — after a shaky, error-riddled start in 1997 — is said to now run virtually problem free, so companies don’t have a compelling reason to abandon the old version just yet.
Warren Shiau, a software analyst at IDC Canada Ltd. in Toronto, suspects companies won’t want to go through the labour-intensive process of updating their software, and hiring the people to do it, so soon.
“There are people out there who’ve finished working all the bugs out (on 8 or 8i) not too long ago, and have it working just perfectly,” says Shiau. “So it becomes very, very hard for the IT manager to justify an upgrade to the chief financial officer or whoever he’s reporting to. Upper management will say, ‘Well, we can afford to wait for Oracle to debug the (new) software themselves, and, while we’re waiting, there’ll probably be more bells and whistles added to 9i later on.’”
Oracle appears to be forcing users to switch to 9i by de-supporting 8 last fall, but, in doing so, runs the risk of alienating its user base.
“My pet peeve is that Oracle brings out (new) versions and de-supports other versions too quickly,” says Pomalis. “I have worked on a number of Oracle projects where the version (being used) was no longer supported, but the effort to go through the upgrade was considered more prohibitive” than utilizing the older version.
Even though 8 might be sufficient for current needs, Shiau expects most Oracle customers to migrate to 9i in early 2003 once version 8 is rendered obsolete by new applications.
“Oracle is no longer just a database company, they’re also an applications and infrastructure company,” he says. “So this is actually where 9i becomes very different than 8, (because) 9i is part of an integrated package of products like their Application Server and Web services initiative. As the functionality of this package increases, that’s when you’ll see people starting to move to it.”