In a session on Oracle relational databases versus NoSQL databases, expert John Kanagaraj, who works for a major e-tailer that can process many millions of transactions per day, said that in the era of big data, companies need to take a closer look at NoSQL database alternatives to traditional relational databases.
"It is something that will be present in the database world, and it competes directly on the low end with Oracle," Kanagaraj said. "One of the big advantages is that it can scale out massively at low cost."
Oracle Collaborate 2014 conference coverage
Don't implement Oracle databases 'just because'
Kanagaraj said the onslaught of the volume and new formats of data has changed the nature of analytics. Traditional tools such as Oracle Database cannot handle the surge in volume and in unstructured forms of information. And there are other issues; for example, the Redo and Undo functions in a traditional relational database management system (RDBMS) don't work for these massive streams of data. "The Internet of Things is a common household word. Your fridge produces data; your phone produces data. To be able to process and store these high amounts of data in a traditional Oracle RDBMS is very, very costly. And there are cases where we can't even scale out."
At the same time, there are downsides, Kanagaraj said. Data consistency can be an issue. Users may confront situations in which their data is inconsistent with another user's view, with reads and writes being out of sync. More important, there can be issues of data loss, he said. Companies need to think about their tolerance for problems with data consistency and data loss, he added, before committing to NoSQL architectures.
For more on the Oracle Collaborate 2014 conference, check out our conference coverage.