According to Juan Loaiza, Oracle’s senior vice president in charge of availability, database management technology has come a long way in terms of being able to offer mainframe-like availability on low-cost servers such as blades.
“A lot of what’s going on is we’re trying to move availability onto what we call the scale-out era,” Loaiza told session attendees. He added that the characteristics of the scale-out era include “commodity building blocks, inherent scaling and redundancy.”
Loaiza said that because of their failover capabilities, Oracle Real Application Clusters are the key to avoiding unplanned downtime caused by server failures.
And in the event of storage failure, Loaiza said Database 11g fine tunes recovery from corrupt blocks and crashed storage arrays. A key aspect of 11g in this area is Automatic Storage Management, which mirrors data, he said.
If the downtime is caused by human error, he said, there’s a good chance that Database 11g’s flashback technologies can help. For example, Flashback Database can “move an entire database back in time” while Flashback Data Archive can automatically store all changes to selected tables in an archive that cannot be modified.
“We architected this so it could flashback against an unlimited amount of time.”
From the sounds of this session, it appears that Oracle is on top of its availability game, but I wonder if DBAs would agree with that. Is Oracle really doing all it can on the availability front? Or are there more features or technologies that you’d like to see added to the system? Post your comments here and we’ll likely contact you for an upcoming news story on the subjects.