Database In-Memory ‘has the best bits of HANA and other rivals’.
Oracle has rubbished claims its £14,000-a-time in-memory database could be switched on by default.
The latest version of Oracle Database 12c includes Database In-Memory, Larry Ellison’s rival to the likes of SAP’s HANA that allegedly speeds up data-processing by 100 times.
The in-memory option went live last week, but database expert Kevin Closson claimed on his blog that the add-on has the potential to be switched on by default.
Oracle has promoted the option as easy to implement, with database architects (DBAs) simply writing three lines of SQL to switch on the feature to speed up data processing for a particular application.
But Closson said DBAs did not need to alter an application’s parameters from default when creating a database to inadvertently incur the £14,000-per-processor fee the in-memory option costs.
All they had to do when creating a new database on 12c was to use the wrong database initialisation parameter file from their repository, he added.
He wrote: "Did I have to jump through hoops and be a raging, bumbling idiot to accidentally trigger usage of this feature? No. There is yet no way I know of to prevent accidental use of the In-Memory Column Store feature/option."
However, an Oracle spokesman denied Closson’s claims, saying specific parameters or commands must be set to trigger the feature.
"In the case of Oracle Database In-Memory, the "inmemory_size" parameter must be set to a non-zero value to enable the In-Memory Colum Store," he said.
"A DBA would need to explicitly alter the default setting in the parameter file to enable the option."
The news comes after Oracle told CBR that making Database In-Memory an add-on gave its customers more freedom, following SAP’s claims that in-memory should come as standard.
Oracle’s director of business development, Duncan Harvey, said: "What we’re trying to give our customers and our users is freedom. We are not forcing them to take any Oracle technology.
"We’ve deliberately released it as an option as opposed to a mandatory architectural change because some customers may have optimised their warehouses or their applications and they’re getting the performance they need. But for the bulk of our audience they’re going to really be able to gain value from it."
Harvey added that in some cases beta customers had sped up data processing by around a thousand times, though did not provide examples.
The technology is expected to be of benefit to hundreds of thousands of Oracle customers, but it comes four years after SAP’s HANA.
Harvey said: "What we’ve done is we’ve looked at what our peers are doing, whether it’s SAP or IBM or whoever, and we’ve really taken a good look at their solutions and we’ve taken the best bits to create our product."