Oracle is introducing a major new release of its NoSQL database - version of 4.0. First unveiled in 2011, the Oracle NoSQL Database is a key-value database that evolved from the company’s acquisition of BerkeleyDB Java Edition, a mature, high-performance embeddable database.
While Oracle is well known for its relational database technology, fundamentally, Oracle focuses on data management in general, and views NoSQL as one of the key products in the data management space from Oracle, said Ashok Joshi, senior director of NoSQL, Berkeley Database, and Database Mobile Server at Oracle.
“Oracle takes a comprehensive view that some customers need relational database technology for their applications, others need NoSQL, and some need Hadoop, or a hybrid,” said Joshi. Today, said Joshi, Oracle is seeing many use cases for NoSQL technology for sensor data capture, IoT, telco call center data management, as well as fraud detection in financial services, online personalized promotions, and advertising. Still, said Joshi, customers have the same expectations from Oracle, no matter the database technology – support for ACID transactions, and enterprise-grade reliability and performance.
Oracle releases quarterly updates to the NoSQL database. Among the 5 key enhancements Oracle is making in NoSQL 4.0, said Joshi, are:
1-Support for “time-to-live”: This is relevant for the IoT and sensor data capture space where typically users want to store 3–6 months’ worth of data, but data that is older is not relevant. This feature enables the option of deciding that data can be automatically aged out of the system or moved offline after a certain set period of time, said Joshi.
2-Predicate pushdown: As part of the larger Oracle data framework, Oracle NoSQL works well with the Oracle Database. It allows a query to be run from Oracle Database against data that is stored in the NoSQL Database, essentially connecting the two repositories together.
“Predicate push down is a very important feature from a performance and usability perspective,” said Joshi. With the predicate pushdown capability that has been added, users can push some of the processing down all the way down to the Oracle NoSQL Database nodes and - since there are a large number of nodes - the computational processing can be distributed across multiple nodes. This allows users to get the subset of results that they are interested in and it is highly parallel, there is distributed computation over multiple processors, and there is also reduced network traffic because it avoids unnecessary data sent across the wire to the database server.
3-NoSQL cloud service: Before the end of 2016, the Oracle NoSQL Cloud Service will be available from the Oracle Cloud. In preparation for that, Oracle has added an export/import feature that allows users to exchange data between two NoSQL DB repositories (for example, on-premise to the cloud) to make it easier and more efficient to move data.
4-SQL access: To address the requests of customers who are familiar with Oracle Database and SQL, with this new release, Oracle is providing the ability to query data in the Oracle NoSQL Database with a SQL API. It provides a subset of SQL, not full SQL capabilities. It allows users to express queries against the NoSQL Database using SQL, a declarative language, as opposed to the API that was previously used.
The combination of predicate pushdown and SQL access will be valuable to customers that query NoSQL data from Oracle Database, said Joshi. “This is a NoSQL database and you can write a SQL query and get results from the NoSQL database, and then join them and do aggregations and have all your existing applications work without even knowing that some of your data is in the NoSQL Database.”
5-Free text search: Oracle NoSQL Database provides secondary indexing so users can access data either through the primary key or secondary indexes such as indexes on last name or ZIP code, but users also want to do free text search. As a result, with this release, Oracle has added the ability using Elasticsearch to do free text search on NoSQL Database for ad hoc searches.
Oracle NoSQL is available as an open source community edition with optional support, and as well as an enterprise edition.
For more information, go here.