michelangelus - Fotolia
Dirk Slama was looking for a nimble database platform to support his company's products for the Internet of Things, and MongoDB fit the bill.M
Slama is the director of business development at Bosch Software Innovations, which is developing software to support big data applications. Pilot projects include taking reads from as many as 1 million cars, and Slama said MongoDB can support "a constantly evolving system."
"With Mongo, agility and flexibility turned out to be really critical benefits," Slama said.
Web applications that spawn unstructured data have driven interest in document-oriented, open source NoSQL databases such as MongoDB, which works with flexible schema. MongoDB Inc. (formerly 10gen), staged its first MongoDB World event in New York City last week, in part to spotlight the database's success in Web apps within commercial enterprises -- and design agility was in the forefront.
IT leaders like Bosch's Slama, as well as several from Citigroup, Verizon and other firms, credited MongoDB for providing agile development and supporting diverse data types, where incumbent relational databases have proved cumbersome.
Establishing mainstream credentials is important to MongoDB, as well as to a pack of startup NoSQL competitors, including Couchbase, Cloudant and others. Changes in the nature of corporate data drive experimentation with the new databases.
"Unstructured data is coming at us at extremely high speeds and in complex ways. We realized that the traditional platforms we were executing on were challenged as we began to look at structured and unstructured data elements together," said Michael Simone, global head of Data Platform Engineering at Citibank's Financial Services Group.
He said MongoDB, initially introduced by "pioneering" developers working on a global financial application, offers faster development time and greater schema flexibility as applications evolve.
DevOps and Mongo
One clear case of agility today is DevOps, an approach to software development where administrative, deployment and development is handled in tandem. Such approaches were fundamental in the rise of Web powerhouses such as Google and Yahoo, and some DevOps fervor is seen in the MongoDB community.
"The lines between the technology areas are fading," said Shivinder Singh, senior systems architect and member of Verizon's Data as a Service Team. He is exploring strategic uses for MongoDB after initial successes with Verizon Intranet applications.
"In the past, there was a team that did sysadmin, a DBA[ database administrator], middleware administrator -- those needs seem to be blurring," he said.
"The keyword is agility: the ability for developers to build and modify applications quickly," said Carl Olofson, analyst at market research company IDC.
That means doing so without being tied to the formal steps of a classic relational database procedure, Olofson continued, describing the common case "where you work with a DBA and a data modeler that have to reconcile any data changes to the overall model."
Olofson said MongoDB is finding use "for managing data on the state of applications: the data the application needs to persist in order to continue to operate in a consistent way." The data can be used later to learn what has happened with the application, he said.
At MongoDB World, Dwight Merriman, company chairman and co-founder, said Agile development support was a major design point when MongoDB, which he launched after founding and selling Web advertising giant DoubleClick, set about to create a model for fast Web application building.
"We were not talking waterfall development," he said." We were talking about lots of iterations -- a release a week, or more often. If you release software every day, is there a schema rev every day?" he asked.
To date, the Agile development and DevOps movements have proved to be something of a dovetail that connects a style of development with new data architecture, as mainstream IT presentations at MongoDB World seemed to indicate. Continually evolving data types could auger more of the same.
Try this quiz to see what you know about NoSQL databases