The high points of Oracle NoSQL Database Release 2.0 include auto-rebalancing, manageability and application programming interface (API) upgrades that address practical deployment and administrative concerns. The auto-rebalancing feature dynamically manages compute and storage capacity to maintain service levels even as processing demands fluctuate as the scale and throughput of data and the number of users varies. A new Web-based management console gives administrators access to all the tools and controls they need to deploy and monitor the database, according to Oracle.
On the development front, Oracle has added a C-based API for those who prefer that language over the existing Java API. A new Large Object API is aimed at handling images, documents and other large objects. An automatic serialization API takes advantage of support for Apache Avro in Release 2.0. A remote procedure call for data serialization, Avro lets you define a schema (using JSON) for the data contained in a record's value. This compact, schema-based data format also eases integration with Hadoop.
[ Want more on Oracle big data moves? Read Oracle Acquires DataRaker For Utility Analytics. ]
Oracle NoSQL Database is based on open-source Berkeley DB, a venerable, decades-old key value store that's widely used in the telecommunications industry. Oracle is hoping its NoSQL spinoff can go head to head with open-source NoSQL databases such as Cassandra, MongoDB, Couchbase and Riak, which have quickly gained momentum in the big data movement in recent years. A free Oracle NoSQL Database Community Edition Release 2.0 is available from the Oracle Web site.
Oracle's latest appliance release, Oracle Big Data Appliance X3-2, has much in common with the Oracle Exadata Database Machine X3-2, which was announced in October at Oracle Open World. Both appliances harness latest-generation 8-core Intel Xeon E5-2600 series of processors. The full-rack appliance packs 18 compute and storage servers offering a total of 648 terabytes of raw storage capacity. The hardware delivers 33% more processing power with 288 CPU cores and 33% more memory per node with 1.1 terabytes of main memory. Power and cooling requirements per terabyte are naturally reduced.
On the software front, the Big Data Appliance includes Cloudera CDH4.1, the latest Hadoop software distribution. Released in June, CDH4.1 was developed to improve NameNode availability and thereby eliminate the single point of failure in a Hadoop cluster. The Big Data Appliance also includes Oracle NoSQL Database Community Edition 2.0, the free version of Oracle's just-released NoSQL database mentioned above.
The Big Data Appliance software bundle also includes upgraded Oracle Big Data Connectors. A SQL Connector for Hadoop Distributed File System, for example, is said to improve query performance of HDFS from Oracle Database by way of Hive tables. The upgraded software also offers transparent access to the Hive Query language from the R statistical programming language. This supports analytic techniques natively in Hadoop, enabling R developers to be more productive by opening up access to the data in Hadoop clusters.
Both the Oracle NoSQL Database Release 2 and the Oracle Big Data Appliance X3-2 are available immediately.
Tech spending is looking up, but IT must focus more on customers and less on internal systems. Also in the new, all-digital Outlook 2013 issue of InformationWeek: Five painless rules for encryption. (Free registration required.)