When the processing job at hand is updating general-ledger data or running queries for analytics and reporting, SQL relational databases are still the likely technology choice for most organizations. But it's a different story with Web applications, particularly ones running in NoSQL-friendly cloud computing environments.
"NoSQL arose from a desire to quickly spread cloud-based applications and Web applications," said IDC analyst Carl Olofson. Relational software became harder to maintain as databases increasingly ran across farms of servers scattered in multiple locations or deployed in the cloud, he added. That created an opportunity for NoSQL database vendors -- and they're cashing in on it.
More on NoSQL database technology and the cloud
Find out what all the buzz about NoSQL software is really about
Try this quiz to see what you know about NoSQL databases
Read why more users are getting on board with cloud databases
Guy Harrison, executive director of research and development for Dell Inc.'s database management tools, said NoSQL databases aren't dependent on cloud environments. But, he added, they "are built with the cloud in mind. They all easily scale elastically, for instance." Harrison, who also has written four books on managing databases, said he expects cloud deployments of NoSQL technologies to increase naturally as more organizations move applications to public clouds.
The strictures of SQL also make it difficult with relational software to modify the data architecture of Web applications and set up different fields and structures for individual data sets -- both common requirements for Web developers. The schema-less or schema-after-the-fact NoSQL database is a means to those ends, and that flexibility is often the chief driver when organizations decide to dip their toes in the NoSQL waters.
"Doing the entire schema in advance is inconvenient for Web applications," said Curt Monash, president of analyst company Monash Research. In addition to complicating modifications down the road, it adds time to the development process -- and on the Web, Monash pointed out, deployment speed is often king.