When MongoDB first emerged about 10 years ago, they were arguably the “barbarians at the gate.” MongoDB’s core premise was that the relational model of data was unnatural and inefficient for modern software developers. They thought that developers would welcome a simpler alternative that was more aligned with object-oriented programming practices and with modern agile development and DevOps.
I’d argue that MongoDB has proved its point over the last 10 years. For a very large segment of the software development community, MongoDB represents the default choice for a database, and certainly, MongoDB is far more popular today than the venerable incumbents of 10 years ago. MongoDB can take pride in the fact that of the top five databases (by DB-engines.com rankings), only they are 21st-century technology. The rest of the top five were all originally developed in the 1980s or 1990s.
However, it is the fate of successful barbarians to become civilized. MongoDB now has to appeal to enterprise decision-makers as well as coalface developers. And, having made their point about the attractiveness of the JSON document model, MongoDB now faces the reality that all of the legacy databases are offering increasing support for JSON data models. The success of “NewSQL” databases such as cockroach DB and snowflake DB represents a resurgence of support for the SQL language. It’s not inconceivable that MongoDB will be challenged in the near future by these NewSQL variants offering “good enough” JSON support.
Therefore, if MongoDB wants to sustain its growth, it can’t be complacent; it needs to be looking to the future. MongoDB is a large company now with many distinguished engineers from a broad set of backgrounds, and a large proportion of these are thinking about the future of MongoDB. However, there is one small group established whose job is exclusively to look at the future of MongoDB, and that is the Sydney-based MongoDB Labs group.
MongoDB Labs is headed up by Michael Cahill, who came to MongoDB via the WiredTiger acquisition of 2014 and whom we met in this column back in March. Michael is a true database veteran and innovator and is particularly well placed to understand the past, present and future of database technology.
The MongoDB labs group was created one year ago with the aim of forming a long-term vision for MongoDB. They are trying to determine what databases will look like in 5 years and beyond and to chart a course for MongoDB to navigate that future. They undertake pure research, track academic work and perform other tasks to identify potential gaps and opportunities in MongoDB capabilities.
As we approach MongoDB.Live 2021, it’s understandable that MongoDB Labs is not in a position to discuss publicly things that are likely to become marquee features in the next release of MongoDB. However, Michael was kind enough to share with me some of the themes that guide his research.
MongoDB currently sits in an important niche within the full life-cycle of data. It’s primarily the recipient of hot transactional data from websites and mobile applications. However, there is an aspiration for MongoDB to play a broader role across the entire data life-cycle. As an example we've seen with MongoDB Atlas Data Lake and ability to work with cold data stored in Amazon S3 buckets. We can speculate that we’ll soon see more support for the management of non-transactional data across a broader range of database use cases.
The MongoDB query language (MQL) and the aggregation framework provided MongoDB with an alternative to SQL. MQL has worked very well for MongoDB and is becoming a de facto standard for MongoDB lookalikes; MQL-like interfaces have been implemented by Amazon and Microsoft, for instance. However, as we noted earlier, SQL is far from dead. Just as the SQL databases are adding support for JSON, it seems likely that we will see increasing interoperability with SQL in MongoDB.
MongoDB has had quite a wild ride over the past 10 years and has succeeded beyond expectations. However, what got them here won’t get them where they want to be in the next 10 years. I can’t wait to see the next phase of MongoDB’s technology evolution.