ORACLE has eyed SAP HANA with the launch of an in-memory option for the its Database 12c software, which promises to speed up the processing of specific workloads and by up to 100 times.
Key for customers is that this has been integrated transparently, enabling existing workloads and applications that use Oracle Database 12c to take advantage of it.
The in-memory option for Oracle's flagship Database 12c platform was first disclosed at the Oracle Openworld show in San Francisco last year. However, while the firm is announcing the technology today, the feature will actually be available as part of release 188.8.131.52 of Oracle Database 12c, due to ship within 60 days.
Oracle is pitching the fact that having in-memory capability integrated with its existing database is a key differentiator, especially against SAP's rival HANA platform, because it offers compatibility with current applications along with existing features for robustness and to guarantee transactional integrity.
im Shetler, vice president of product management at Oracle said, "It is completely transparent to implement for existing applications that work with the Oracle database. This is different from all the other [in-memory] offerings in the market today that either require changes to applications or limit functionality to a subset of full database functionality."
Singling out SAP HANA, Shetler said that it is effectively an in-memory data store that is very fast, "but they are still trying to complete the rest of the database functionality around it," he claimed.
"With Oracle, all of the features of the database are available, all of the applications that exist today, including third-party applications, custom-written applications, they will all work out of the box with the Oracle Database In-Memory option, so we think that's really huge benefit to enabling companies to become real-time enterprises," he added.
Another area where Oracle claims an advantage is in the size of the database that can be used with its in-memory technology. Users can allocate a region of memory to hold the database and specify which data they want to go into that region, whether this is an entire table, a portion of a table, or a subset of the columns in a table.
"It's very typical for an analytics application to only look at a small number of columns. A report might have only 10 data items out of a table that might have 500 columns, so it's important to conserve space to be able to identify precisely just the data that needs to be in memory, and we give you the ability to do that," Shetler said.
The in-memory option also works with Oracle's Real Application Clusters (RAC) feature, which allows a single database to be spread across multiple servers, enabling a much larger data set to be accommodated.
On Oracle's Exadata systems, the in-memory option allows the data to be spread across memory, flash and disk and transparently accessed, according to Oracle, so once more the entire data set does not have to be present in memory at the same time.
Oracle said many of its application teams have been working to incorporate the in-memory option into their software, and from its experience, it believes developers will see a performance boost without changing their applications, but an even greater boost if they update their code to take best advantage of in-memory processing.
"Some applications such as Oracle PeopleSoft, financial applications or the E-Business Suite have seen several hundred and up to a thousand times speed-up through a combination of adding the in-memory option and doing some restructuring of their internal algorithms," Shetler claimed.
Oracle is also extending its partner programme to certify applications on Oracle Database 12c with the in-memory option.
The Oracle in-memory option will be available on all platforms where Oracle Database 12c is currently supported, but availability details have not yet been announced.. µ