Follow Datanami:
June 9, 2020

MongoDB Steps Up Game with MongoDB Cloud

(Nomad_Soul/Shutterstock)

First there was Atlas, MongoDB’s hosted database-as-a-service (DBaaS) offering for public clouds. Now the company is launching MongoDB Cloud, an amalgamation of a range of MongoDB products and services that includes Atlas, the ability to query S3 data, synchronization with mobile databases, and more.

June is traditionally the month that MongoDB holds its annual user conference, MongoDB World, usually in its hometown of New York City (except for that 2017 show in Chicago). With COVID-19 still making the rounds, the company was forced to hold a virtual version of MongoDB World, called MongoDB.live, which is taking place today and tomorrow at www.mongodb.com/world.

As part of its virtual MongoDB Live activities, the company is making a slew of announcements, including MongoDB Cloud, Atlas Data Lake, Atlas Search, MongoDB Realm, and version 4.4 of the MongoDB database.

MongoDB Cloud is the new umbrella brand that the company will use to refer to all of its cloud offerings, including but not limited to Atlas, says MongoDB Chief Customer Officer Richard Kreuter. “With MongoDB Cloud, we’re adding or GA-ing a number of critical services that are complementary to the core operational database,” he says.

At the core of MongoDB Cloud is Atlas, which the company launched way back in 2016, becoming one of the first (if not the first) managed NoSQL databases available on the cloud. Since then, the DBaaS offering has grown considerably, and last week we learned that it accounted for 42% of the company’s total revenue in the first quarter of fiscal year 2021, or almost $55 million.

Another offering under the MongoDB Cloud umbrella is Atlas Data Lake, which was introduced as a beta offering last year and is now generally available. Atlas Data Lake gives customers the ability to query data that’s stored in AWS S3 buckets using the MongoDB query language. This gives customers more flexibility in how they store and query data, Kreuter says.

“You can put data in S3 in a number of common arbitrary file formats, but you can still query it and work with it as if it were a running MongDB database,” he tells Datanami. “It speaks MongoDB network protocol, wire protocol, and can answer MongoDB queries over data that’s stored in that way.”

The company is also tightening the integration between Atlas Data Lake and the core database to enable online archiving from Atlas into the data lake. “This means that you can set things up so that, as data as data gets older, it’s automatically archived off to S3” based on time stamps, Kreuter says.

It’s also unveiling a federated query capability that allows MongoDB users to query both the Atlas and S3 store using a single query. That will simplify life for developers by eliminating the need for them to look in multiple locations for the data, Kreuter says.

“The Atlas data lake and the database spoke the same query language, but there would be two different end points that you have to keep track of. [You would have to say] ‘Do I want to query data over here or over there?’ Now with federated queries, you don’t have to do that. We do that for you.”

Atlas Search will also go GA. Unveiled last year as a beta, Atlas Search brings the power of Apache Lucene and its full-text indexing capability to Atlas. According to Kreuter, it gives users Google-like search capabilities for any data stored in the hosted DBaaS. “It gives you a very  easy way to leverage MongoDB but with natural language-type fuzzy matching, synonyms–all that good stuff,” Kreuter says.

Last year, MongoDB bought an embedded database company called Realm. Today, MongoDB announced that it has fully integrated the popular database into the MongoDB Cloud, under the name MongoDB Realm. Customers can benefit by having their MongoDB Realm data automatically synchronized and backed up from the mobile device into MongoDB Cloud.

(Valery-Brozhinsky/Shutterstock)

“It’s really making it easier for developers to work with data on the mobile device, to have that data be visible, accessible, and workable for them in MongoDB on the backend with no extra work needed,” Kreuter says. “If they’re using Realm, they can just point Realm at that endpoint for synchronization purposes.”

Lastly, MongoDB will be unveiling new capabilities in version 4.4 of its NoSQL data store. These, of course, will be available to all MongoDB users, including those who run the database on-premise, in a private or public cloud, and those who get their MongoDB via Atlas.

The new union operator will bolster analytics in MongoDB, Kreuter says. “It allows you to write certain kinds of aggregation and analytic workloads in a simpler way,” he says. “For example you can say, I want one query that gets the result of multiple queries, all in one go. This allows for richer and faster analytics. And also, by giving people richer and faster analytics, it reduces the need for ETLing data from an operational location.”

Another new update, refinable shard keys, gives customers more control over how they organize data on MongoDB clusters. “This makes certain kind of scale-out operations easier,” Kreuter says. “You can make your shard-key definitions more precise, let’s say, than it used to be.”

For example, if a MongoDB user started out with a course-grained sharing approach that separated customers in different shards according to what country the live in, the refinable shard key functionality will allow users to get more granular with their data organizations by adding county or city codes, Kreuter says. The database will automatically rebalance data based on the keys used.

Finally, the new hedged read capability will give MongoDB users more control over how data is returned. The database was developed to favor strong data consistency, and to return the most current data for a given query, by default. However, that mode didn’t fit all customer use cases, Kreuter says.

“Some customers have told us they prefer they would rather read from whoever can respond fastest, even if that node happens to be a few millisecond behind current,” the Chief Customer Officer explains. “So what the hedge read capability delivers is that lower-latency response time from the database for the purpose of providing a higher quality user experience to the end user of the application, even when some nodes might not be working optimally and when some nodes might be further away than others.”

MongoDB, which claims more than 17,000 customer around the world, will be hosting keynotes and other presentation during MongoDB.live.  You can register for the virtual event here.

Related Items:

A MongoDB Secret Weapon: Aggregation Pipelinev

MongoDB Takes Another Big Step Into Clouds

MongoDB Struts Its NoSQL Stuff in NYC

Datanami