During the first database wars in the early 1980s, it wasn’t immediately obvious that the relational database model was superior to existing models. But over time, this model won due the power of SQL programming and ability to simply structure critical business data for back-office automation, e-business, and web-sites. As data growth and transaction loads increased, bigger and more powerful servers were developed to “scale-up” database performance and database vendors optimized their products for these platforms. Parallel database technologies operating across many servers then emerged to address more complex analytic queries and large data sets that could not economically be done with large scale-up servers.
As the volume, velocity and variety of data increased so did the connectivity of this data to users and devices. All this created a breakdown of the traditional scale-up database and complementary approaches to the relational database. Three new scale-out technologies are leading the newest database revolution: NoSQL, distributed SQL and Hadoop. Each of these takes advantage of pay-as-you grow building blocks of cloud computing and delivers massive linear scale of both data and queries on the data. Together they provide a new foundation for companies building the new generation of innovative data-centric applications.
The graphic puts this shift in the context of prior historical turning points and the rapidly evolving near future.