I liked this post from the Guardian blog about four predictions for Big Data in 2014. It was played pretty safe; and I expect this prediction to be just right. In addition, I think, we would see divergence of the Big Data topic by the fields: when people learn and become tired of technological terms, like Hadoop, noSQL, they will start talking about specialized applications and approaches. Then it will become really fascinating.
Aerospike, the company developing an in-memory noSQL database, offers seven predictions for Big Data in 2014 that play more on a technological side. It is no surprise that one of the predictions is that “database architectures with operational in-memory NoSQL databases in front of analytic data warehouse will become standard.” Could not notice that this is an exact solution I proposed in my recently pending NSF application for further development of our IntegromeDB search engine and integrated database.
Aerospike also predicts “a greater use of Amazon’s platform-as-a-service (PaaS) for analytics, with real-time delivery of applications from cloud systems that offer higher performance and reliability.” Indeed, the estimate is that “between now and 2017 the market for PaaS solutions will increase by 30 percent annually, to reach a turnover of $14 billion by 2017.” And another sound prediction is that “at least one major national retailer will begin offering a free API access for their data to drive development of mobile apps for shopping recommendations and delivery services.” Databases with no API will eventually become useless.