The official launch of Oracle Database 11g happened on Wednesday, July 11 in New York.
But the build up to this pivotal event began more than eight months ago.
It started when Chuck Rozwat, executive vice-president of Oracle server technologies announced the beta of the 11g at Oracle’s annual user conference in San Francisco last October.
I was at Rozwat’s keynote at the Moscone Centre, and the message from him and other Oracle execs was clear – 11g represents a watershed in database technology.
With 482 new features, this database – we were told – would blow the competition out of the water.
In the past several months, scores of “experts” have participated in Oracle 11g’s beta program and offered wide-ranging perspectives, opinions and advice.
Their collective commentary has tended to focus on features of the soon-to-be released database that are perceived to have the greatest business impact — such as:
Data retention and compression
11g, we hear, will feature new “compression technology” that could reduce customers’ storage demands by two-thirds, by enhancing their ability to store data – especially unstructured data – faster than traditional file systems.
This is being promoted as a key capability, given the increased storage and data retention pressures companies face today, including the burgeoning growth of rich media and increasingly stringent compliance regulations, requiring companies to retain documents – including e-mails and instant messages – for longer periods.
These are strong enough drivers to consider a system such as Oracle 11g, according to Oracle expert Rich Niemiec.
Niemiec, former president of the International Oracle Users Group, has authored several books on Oracle technology, including Oracle 10g Performance and Tips.
In an interview published in Oracle’s Database Insider newsletter he says the swift pace of data growth and change calls for “systems that give us a way to visualize and manage huge data loads” and “help us meet change with confidence.”
11g, he says, fulfills this role with its (expected) “workload capture and replay” capability.
This feature will be great for managing change and tuning systems, Niemiec says.
“People are constantly making changes in their data centers: implementing migrations and upgrades, or changing hardware, operating system, and applications.”
11g is expected to offer a menu-driven way of capturing a system’s workload over time and then replaying it precisely in the same manner as it was captured.
This “workload capture and replay” feature, Niemiec says, allows you to ensure consistent or better behavior when you make changes.
“You can capture workload for a minute, an hour, or for several days. Then, when you make changes in your system, you can test your real day-to-day workloads against the new system [and] make changes with confidence.”
The feature, he says, is also great for database fine tuning, as each time you make a change, you will know precisely how it affects various system parameters.
“The system would tell you, among other things, the number of block reads that it’s going to change, or how the CPU is going to change, and [would] graph it out to show you the before and after for this whole group of SQL statements.”
This workload capture capability enables the entire production workload to be brought into the test environment, notes Leng Tan, vice-president of database manageability and diagnosibility at Oracle in a podcast.
She said this feature is particularly useful to database administrators (DBAs) who to frequently have to update software, change the operating system or introduce configuration changes because the applications require it.
“Traditionally, when you introduced these changes into production there was a risk of instability. To mitigate those risks, we’re introducing functionality within 11g to make the testing these changes a lot more realistic within the test environment, before they are introduced into the production system.”
Once captured, she said, the production workload can be replayed in the test environment with the same concurrency and critical timing.
“You then introduce your change – [say] migrating from a non-RAC to a RAC system or from a non-Linux to Linux.”
She said this ability to test a change in the test environment – but with the production workload is unique to 11g. “None of the other solutions out there has it; we’ve talked to quite a few beta customers about this capability it’s been really well received.”
Other 11g “change assurance” capabilities – in addition to “workload capture and display” – are discussed by Mark Townsend, Oracle vice-president, database product management in a presentation.
They include the ability to: detect and tune performance changes, package incidents for support, and do hot fixes and online application upgrades.
Townsend notes that large, business-critical applications can often remain unavailable for hours while an upgrade is installed.
Oracle Database 11g, he says, allows online application upgrades to be done while the application remains continuously available.
“The pre-upgrade application and the post-upgrade application can be used at the same time [and] each end-user session is rolled over on its own schedule.”
Support for online hot patching is another key 11g “change management” capability, according to Townsend’s presentation.
While Oracle already supports rolling patches between nodes in a cluster, he says the forthcoming release adds online patching of a running Oracle executable (single instance or cluster) with no downtime.
Another 11g feature that could drive definite business benefits is its “learning optimizer” feature, says an Oracle commentator.
The query optimizer exemplifies the self-management capabilities of 11g, according to Iggy Fernandez, editor of the NoCOUG Journal, the official newsletter for the Northern California Oracle Users Group.
“When relational databases replaced hierarchical and network databases in the 1980s, the promise was that programmers would no longer need to optimize their queries by hand,” said Fernandez.
He said while no database has “completely realized” that goal, 11g is comes close. “The query optimizer simply learns from its mistakes – in fact, it can stop a query that is already in progress and try a different approach!”
“Performance tuning is the biggest component of database administration effort today, and the learning optimizer could produce tremendous labor savings if it works as advertised,” Fernandez said.
The XML DB – much ado about nothing?
Among the more controversial features expected in Oracle Database 11g are new XML-related upgrades – including a new XML binary type and a new XML index.
At least one expert doesn’t have much use for the upgrades.
Brian Peaseland is an independent Oracle database consultant who helps companies and individuals get more out of Oracle’s database management system
“To me, most of the XML stuff inside the database is not important at all,” says Peaseland in an interview in <a href="http://search