MongoDB grows with Atlas Data Lake and mobile services

Atlas Data Lake expands MongoDB Cloud services

MongoDB recognizes that modern application data architecture requires more than just a core database, said Sahir Azam, chief product officer.

Atlas Data Lake takes the MongoDB document-oriented query language and enables developers to run analytics queries on data that may not have originated in a MongoDB database, Azam said. With the data lake service, users can now do federated queries across data stored both in the MongoDB Atlas cloud database as well as from the AWS S3 cloud storage service.

Data Lake is an expansion of their [MongoDB's] capability to enable users to collect and query data from multiple sources.

A key part of the Atlas Data Lake service is an integration that enables MongoDB users to more easily archive data. Azam noted that certain types of applications, including data from internet of things sensors, can grow to petabytes in size, which is not economical to store inside of a transactional database like MongoDB.

What typically happened in the past with large data sets is that users copied the data out of MongoDB and then stored it in AWS S3. The challenge, however, has been that the data wasn't easily accessible to query from MongoDB.

"What we do now is fully automate the process so customers can set a rule saying, you know, 'This is the data I want to be automatically archived,'" Azam said. "We handle the archiving, but most importantly, we enable that data to still be query-able by the core system, without a change to the application at all."

Bringing MongoDB to the mobile realm

The new MongoDB Realm service, now generally available, expands on the open source Realm database, enabling developers to run an optimized mobile database that can synchronize data with the MongoDB cloud.

MongoDB Realm gives developers a way to build mobile applications that require device synchronization with the cloud, which can have multiple applications. For example, a mobile app user might not always have connectivity while travelling by subway. But with local on-device database the app will still work, and then once the user regains connectivity, the app can be synchronized with the cloud for updates.

"We're not taking a server-side back-end database and trying to pretend it's a mobile database, and conversely, we're not taking a mobile database and trying to make it a server-side database," Azam said. "We're combining forces with the best of both."

Updated MongoDB platform improves performance and scalability

While MongoDB is expanding its catalog with the Atlas Data Lake and mobile database services, it's also updating its core open source database.

MongoDB 4.4 introduces a feature known as refinable shard keys which provides a way for users to better manage how data is distributed across a MongoDB database cluster. Database sharding is an approach that slices up or "shards" data into multiple units that are spread across a cluster. Over time as a database grows, users often need to adjust how shards are distributed, Azam said.

"If you need to refine the way data is actually distributed you can now change that in the database and we will manage the redistribution of that data seamlessly in a non-disruptive way," Azam explained. "So it makes it a lot easier to adapt the database as the scale requirements of an application and data partitioning change over time."

Another new feature in MongoDB is hedged reads, which can help to improve database lookup speed. Hedged reads routes database read requests across a MongoDB cluster in a more optimized approach.

"Fundamentally, we're trying to simplify the way that developers work with data so they can focus more of their time on building great applications," Azam said.