Following its success with Amazon Web Services, MongoDB’s namesake database can now run on cloud services from Microsoft and Google.
“Amazon continues to be the dominant place to run applications in the cloud, but there is a lot of interest, especially from Microsoft shops, in running MongoDB on Azure. And there is significant interest from developers in what Google has to offer,” said Matt Asay, MongoDB vice president of marketing and corporate strategy.
All three cloud-hosting services “understand developers and build the tools and infrastructure to make developers really happy,” Asay said.
The company, estimates that 40 percent of MongoDB installations run on Amazon, which translates into “tens of thousands” of instances, Asay said. Rackspace, MongoLabs, MongoHQ, and ObjectRocket also offer MongoDB as a service, among others.
MongoDB is now available, in preview mode, on the Google Compute Engine as a managed virtual machine. This means developers can deploy MongoDB directly from within Google Cloud through a “click of a button,” and not have to worry about managing any of the supporting software, Asay said.
Pricing for the Google service has not been established.
Microsoft Azure also now offers MongoDB as a managed service as well, as an add-on in the Microsoft Azure Store.
While the Google service is primarily for developers, the Microsoft service is better for production duties, given that it is based on the full MongoDB Enterprise commercial version of the software and includes support from MongoDB itself, Asay said. The Azure service also features associated replication, backup, monitoring and support services.
The cost of running the Azure MongoDB service starts at US$400 a month, for a three-node instance. Each node, running on a Linux virtual machine, comes with 1.75GB of RAM and 40GB of storage.
The services were announced at the MongoDB user conference Tuesday in New York City.
MongoDB will also be showcasing enterprise users at the conference, including clothing retailer the Gap; the financial services company Citigroup; investment banker Goldman Sachs; and agricultural machinery maker John Deere.
The MongoDB database has found favor with developers due in large part because of how easily it works with Web applications, particularly those encoding data in the JSON (JavaScript Object Notation) format, which MongoDB can natively digest without the need for connectors or manual data remapping chores.
Dynamic schemas are another advantage MongoDB offers, Asay said. This feature allows developers to easily add another column into an existing database table, without the need to reformat the data already captured before the column was added.
While dynamic schemas may seem like a minor feature, they can be a huge time-saver, especially for organizations that have to frequently make changes to their databases.
Classified advertising service Craigslist found this feature handy, for instance, when adding a column to housing listings to say whether pets are allowed in an apartment. Such a feature would take months to add to a standard relational database.
“It’s a simple thing. It probably took a developer 15 minutes to write that code. But it would take three months to comb through its terabytes of data to push out the columns, row by row, to have that one binary value, yes or no,” Asay said.
At the conference, Facebook subsidiary Parse, a mobile analytics firm acquired by Facebook in April, will also unveil the open-source release of a number of tools related to MongoDB.
Parse has opened sourced Flashback, which developers can use to test and tune performance of their MongoDB databases, and Dvara, which can extend the number of active servers a MongoDB database can use.
Open-source data analysis vendor Pentaho also announced that the new version of its flagship product, Pentaho 5.1, comes with the capability to run code-free analytics directly on MongoDB data stores.