There's a revolution under way in the way documents are managed. Open-source platforms are rapidly replacing a host of proprietary systems as part of the rise of NoSQL databases in the enterprise.
Most enterprise IT organizations would generally prefer not to support additional database formats, but developers are demanding it. They need lightweight, hierarchical frameworks that are not nearly as complex or expensive to manage as the traditional SQL databases.
For example, the MongoDB open-source document database is being deployed at organizations such as Goldman Sachs and MetLife. "MongoDB fills the gap between file systems and relational databases that don't offer a lot of flexibility," Warren Finnerty, vice president of Goldman Sachs, told us. "It turns out, however, that the space between file systems and relational databases is pretty big."
His firm is using NOSQL to build everything from secure file-sharing services for users of mobile computing applications to social pipelines that will combine internal and external news feeds in one stream.
MetLife has announced that is has built a customer service application and a developer recruitment application based on MongoDB.
Merv Adrian, an industry analyst with Gartner, told us that document databases such as MongoDB, CouchDB, and MarkLogic are filling a void once occupied by departmental databases such as dBase. "We're seeing the emergence of a whole range of database and data stores. In the case of MongoDB, it's ideally suited for applications that require hierarchical structure of documents."
The new options are much more scalable than what they are replacing. But Finnerty said NoSQL was no easy sell within the financial services firm's IT organization. It was not overly excited about having to support yet another database platform. However, the gap between file systems that provide no real structure for building applications and traditional SQL databases that developers often find difficult to use was too wide to ignore.
Open-source NoSQL gave the firm an inexpensive way to build those applications without having to commit to proprietary formats. "The firm is very hesitant to lock into any one framework," Finnerty said. "We would only want to look into something that is open."
It's hard to say how big a void document databases ultimately will fill. Adrian expects it to be quite substantial. Developers are looking for more agile platforms that help them build applications quickly.
Enterprise adoption of these database formats is likely to change the role of the database administrator. Adrian said that, instead of administering a single type of SQL database, DBAs will morph into data management architects. Each DBA probably will end up managing more varied types of database instances than ever, but advances in IT automation should make the cost of managing all those database instances substantially lower than it is today.
Of course, a number of security, management, and governance issues still need to be addressed with NoSQL databases. But with more developers starting to vote with their code, it's only a matter of time before document databases become a standard data management platform at more enterprises.