MongoDB in the Oracle Cloud? New Silo-Busting API Makes It Possible

Oracle introduces a new way to integrate two of the industry's leading database platforms

Data management sometimes feels like an almost religious decision: Do you believe in relational databases and SQL development? Or a document database and JSON?

Oracle is making that choice easier with its new API for MongoDB. The Oracle Database API for MongoDB makes is possible to develop and run MongoDB apps, using MongoDB tools and drivers, on Oracle’s Autonomous Database cloud service.

In other words, Oracle is bridging its own relational model and MongoDB’s document model so customers can have one foot in both camps. “The API allows us to give MongoDB applications compatibility to Oracle databases—and it can expose relational data as MongoDB collections,” says Gerald Venzl, distinguished product manager for Oracle Cloud. Collections are how documents are organized and stored in MongoDB.

The API provides connective tissue between two of the database industry’s leading platforms. Oracle and MongoDB are both in the Cloud Database Report’s Top 20. And both databases score high in DB-Engine’s popularity ranking: Oracle #1 and MongoDB #5.

Integration and interoperability are always valuable objectives in the balkanized world of data. The Oracle Database API for MongoDB provides a level of compatibility between MongoDB and several versions of Oracle’s Autonomous Database—Oracle’s JSON Database, as well as its Transaction Processing and Data Warehousing versions.

I say “level of compatibility” because it’s not 100%. More on that shortly.

Like MongoDB in the Oracle Cloud

MongoDB is available in Google Cloud, but not Oracle Cloud. The API gets closer to that by allowing MongoDB workloads to run in the Oracle Cloud Infrastructure (OCI), albeit on Oracle’s Autonomous Database.

For many users, the natural starting point will be the Oracle Autonomous JSON Database, which was introduced in August of 2020 and, like MongoDB, is a JSON document store.

Oracle claims a few advantages in side-by-side comparison to MongoDB’s Atlas database-as-a-service, including SQL access to JSON documents and a lower price point.

Oracle-MongoDB compatibility gets even more interesting when you think about the integration possibilities of MongoDB apps with Oracle data warehousing or transaction processing, along with Oracle’s ACID transactions, SQL analytics, security, and governance.  

There are a few caveats.  For one thing, Oracle has built its compatibility around MongoDB version 4.2, not the latest and greatest 5.1. That’s not unexpected, given the engineering work involved. For example, AWS’s Amazon DocumentDB database is compatible with MongoDB 4.0, so Oracle is a step ahead in that regard.

And Oracle isn’t claiming 100% feature parity.  At this point, the Oracle API does not support MongoDB aggregation pipelines, which are the step-by-step operations involved in returning query results. You can read about Aggregation Pipelines on the MongoDB site.

However, Oracle says that once MongoDB document collections are brought over to the Autonomous JSON Database, customers can forget about aggregation pipelines and instead use the more powerful and easier to use SQL language for aggregations and analytics.

How will MongoDB respond?

I don’t expect MongoDB to have nice things to say about Oracle treading on its turf. At AWS re:Invent, MongoDB CTO Mark Porter gave a presentation titled, “Is Relational the New COBOL?” That kind of sums up MongoDB’s point of view, and my guess is that MongoDB will continue to hammer on the legacy theme with Oracle.

[It’s worth noting that Porter worked on the technical staff at Oracle in the 1990s. And BTW, I worked at both Oracle and (briefly) MongoDB.]

But MongoDB—which bills itself as “the leading modern, general purpose database platform”—may be overestimating itself and underestimating Oracle. For anyone who’s counting, Oracle had more than 10X the cloud revenue of MongoDB in the companies’ most recent quarters: $2.7 billion for Oracle IaaS + SaaS compared to $217.9 million for MongoDB’s Atlas subscriptions.

Ultimately, customers of both companies are the winners here. Because any advances in data and applications compatibility are helpful in today’s hybrid, multi-cloud data environments. As Gerald Venzl says, without the kind of integration enabled by Oracle’s MongoDB API, “you just continue building data silos.”