BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
It took a few years, but a decadelong march toward cloud computing has now gotten the complete attention of Oracle's technical leadership. While enterprise applications were earliest to experience cloud changes, SQL relational databases now follow their lead. That means new architectural angles in the Oracle cloud database, according to author and database expert Guy Harrison.
It wasn't that developers at places like Google and Facebook disliked SQL, said Harrison, a Melbourne, Australia-based independent consultant, who spoke with SearchOracle at Oracle OpenWorld 2016 in San Francisco. What they did dislike, however, were relational technologies' resolute adherence to immediate consistency and durability in the database system.
As a result, "eventual consistency" became more common in the cloud database. For the emerging Oracle cloud database, some of that approach may be in order. In Oracle 12c Release 2, database sharding, a technique that has been prominently used in cloud computing, becomes part of the mix. Sharding supports extensive scaling of database clusters, but it delivers different levels of consistency than many longtime Oracle developers are used to.
The new database, consistency, sharding, JSON and other issues are discussed by Harrison in Next Generation Databases, recently released by Apress. The book spans the first days of relational development, through to the early cloud era and its demands for new data architecture, to the present, where young databases like MongoDB and Cassandra vie, in some cases, with the relational alternatives of Oracle and others.
In the video, Harrison shares his view on the general trend toward cloud adoption. He says it is now seen as a favorable alternative to on-premises implementations, especially for brand-new applications. Building large-scale infrastructure on premises yourself, when it can be obtained on a subscription basis from cloud vendors, is a virtual nonstarter, he indicated.