How Big Data SQL will let Oracle customers get on with Hadoop.
Oracle hopes to tackle data silos by allowing customers to store and call up data from SQL, NoSQL and Hadoop databases with its latest release today.
Big Data SQL should integrate all types of structured and unstructured data from Hadoop, relational and NoSQL databases, according to the firm, allowing customers to run one query across their Oracle database and all others at the same time.
The tool is purported to minimise data movement, leading to faster analysis across all data types as well as reducing security risks when data is in flight – Oracle Database security and encryption can be extended to data stored in Hadoop and NoSQL as well.
Neil Mendelson, VP of big data and advanced analytics, told CBR yesterday that customers want to use Hadoop, but don’t want to create another data silo.
He said: "You’ve seen over the last six months a real clamouring for SQL on Hadoop and SQL on NoSQL.
"Our customers are looking for new insights by pulling together all sorts of existing and new types of data. [But] they’re having difficulties in terms of integrating this new Hadoop world with the rest of the information they have in their information architecture."
Big Data SQL therefore allows them to run one SQL query across Hadoop, NoSQL and Oracle Database.
Mendelson added: "What we’ve done is to take some of Oracle’s key intellectual property from our Exadata machine, and move that natively onto Hadoop architecture."
He was referring to Exadata’s Smart Scan technology, which is being used on Hadoop to scan and process data at the point of storage, therefore increasing the speed at which queries run.
Big Data SQL will be released in the autumn, with production estimated for August or September – but customers will need Oracle Database 12c to use the tool.
Mendelson did not discuss pricing but said it would be priced "competitively".