By Jeff Erickson
If you’ve been in the database game as long as Kerry Osborne, you’ve gotten used to upheavals due to leaps in the technology. Osborne has built complex and ambitious database architectures for banks and telecoms in the midst of tech trends that have swung “from integrated solutions to best-of-breed and back again.” Meanwhile, says the Oracle ACE and Oak Table member, his database architectures have had to adjust to the rise of the internet and open source software; exponential data growth; and, now, the growth of cloud-based infrastructure and databases.
At every turn, there have been new opportunities. “I just had to be willing to change my focus a little,” he says. When Oracle launched Oracle Exadata Database Machine, it effectively undercut his business—building complex multinode Oracle Real Application Clusters solutions—with a preintegrated package that “did what we did, but better and cheaper.” In response, Osborne had to move his work up the value chain a little and stop worrying about “the low-level integrated components, getting the network setup correctly, and all that stuff.”
The arrival of Oracle Autonomous Database, which uses artificial intelligence in Oracle’s second-generation cloud infrastructure and in Oracle Database 18c to deploy, tune, patch, and secure the database with no human intervention, is causing a similar disruption for many of the DBAs Osborne talks to. “I’ve had lots of people asking me, ‘What’s going on with what I’m doing as a DBA?’” His advice? Roll with it.
“If you’re building databases, if you’re patching databases, or if you’re adding storage to databases, those jobs over time are going to go away,” advises Osborne. But not right away, he says, because there will be a long tail of traditional DBA work as on-premises databases continue to play a role for some companies. For younger DBAs who are doing this kind of work, however, he says “those infrastructure jobs will dwindle.”
But not all traditional DBA work is going away. “If you’re doing performance optimization, architecture, and design work, and if you’re solving hard problems, those skills, as well as communication skills, are things where humans have the advantage over AI,” Osborne observes.
One opportunity he sees emerging is to help AI do its job better—a tactic he learned when Oracle first came out with the cost-based optimizer, he says. The optimizer uses query history and other information to choose the best way to retrieve the data. “It was a way to let the database decide for itself how to do some of the things that database administrators were already doing,” he says.
“Skills then shifted to focus on how the cost-based optimizer works. The folks who really had a good understanding of the way the optimizer did its calculations and better understood why it was doing what it was doing could build the systems or tune systems to help the optimizer do the right things,” he says. “Over time the optimizer got better and better, and DBAs had to do that less than less.”
Osborne sees the same dynamic at play with the newer machine learning capabilities in Oracle Autonomous Database. In your database design and modeling work, “Be the one to understand how the current optimizer and other, newer machine learning processes work,” he says—whether the AI is identifying and remediating threats to the data, patching databases in the background, or predicting user behavior. Figure out how to “get the most out of those processes now and adjust as they get better and better at what they do,” he advises.
Do that, and “If you’re not doing public speaking, go sign up for a Toastmasters class or something where you can start practicing your public speaking. It doesn’t help to have the best idea in the world if you can’t communicate,” Osborne concludes. Plus, “that’s a human thing that a computer can’t ever do.”
LEARN more about Oracle Autonomous Database.
Illustration by Wes Rowell