Oracle has kicked off its Oracle OpenWorld 2013 conference in San Francisco by unveiling in-memory processing for its Oracle Database 12c product, as well as new hardware to run it on, and a new cloud-based database backup solution.
Taking the stage today for the conference’s first major keynote, Oracle CEO Larry Ellison claimed there were currently only two alternatives when storing data in databases – storing it in columns or, as “various researchers” had proposed – storing data in rows.
“Well, we have a better idea,” said Ellison. “What if we store the data in both formats simultaneously? Keep it in row format, but add to that row format the same data rearranged in column format, and while we’re at it, let’s put the column and rows into memory.
Ellison claimed that data can now be processed over 100 times faster, and the company can also now do away with analytic indexing, as in-memory makes it effectively redundant.
“Maintaining those analytic indexes is very expensive, and it slows down transaction processing. Let’s get rid of them,” said the CEO. “Let’s just throw them all away and replace them with the in-memory column storage. Its purpose was to make queries run faster, but now we can run them 100 times faster anyway.”
Ellison took pains to impress upon delegates that using the in-memory option for Database 12c is an entirely voluntary act on the part of the customer, however he was also keen to stress that switching it on would be a seamless, risk-free transition.
“There are no changes to SQL, there are no changes to your applications, there are no restrictions. Whatever works on your networks today runs with the in-memory setting turned on. There is no unloading and loading of data,” said Ellison. “Every one of your applications – every one you wrote, every one you bought, runs without a single change to the application.”
Ellison went on to say that “all of this technology is cloud-ready – ideal for the cloud, with the unique proposition that everything runs 100 times faster simply by hitting a switch”.
Next, Ellison revealed the M6-32 Big Memory Machine. Directly superseding Oracle’s existing M5-32, the new server hardware apparently has twice the cores of its predecessor and is capable of silicon switching rather than employing fibre optic, making the M6 the “fastest machine in the world for databases stored in memory,” said Ellison, as well as “less than a third the cost of IBM’s” nearest specced machine, said Ellison, before joking that an IBM partner in the room looked less than happy at the news.
Finally, Ellison announced Oracle Database Backup, Logging, Recovery Appliances – a cloud-based database backup and restoration system. According to Ellison, previous database backup solutions were “designed for file systems, not databases”, and this made recovery less agile.
By monitoring change logs on databases, the new solution will, according to the CEO, be able to recover data “exactly to the last transaction” quickly and easily. The data can also be backed up in Oracle’s public cloud, “but this is strictly your choice,” said Ellison.
Does Oracle’s in-memory announcement finally give the company what it needs to compete with SAP’s HANA, or has the company – from which Ellison took an 18 per cent pay cut for fiscal year 2013 – already missed the boat with in-memory processing?
Join Computing at Oracle OpenWorld 2013 for the rest of the week’s announcements, customer interviews and other sessions of interest.