MongoDB: From jokes to juggernaut

Once I rejoined MongoDB in 2021, I acquired to listen to all of the previous jokes rehashed. You already know, about MongoDB being “net scale,” about dropping information, about solely being ultimately constant, and so forth. The net scale video is humorous; the opposite statements have largely been incorrect for the reason that day they have been written. For instance, MongoDB has all the time been strongly constant. The contentions that had some semblance of fact have change into outdated with every MongoDB launch. As Senior Developer Advocate Mark Smith notes, “Every part you understand about MongoDB is incorrect.”

After all I’d say that. I work for MongoDB, in any case.

Even so, I feel it’s worthwhile to gut-check our assumptions. For instance, for years we have been advised that enterprises couldn’t substitute Oracle Database or SQL Server with PostgreSQL. For a lot of workloads, that’s merely not true as we speak and nearly definitely wasn’t as “true” earlier than as some would have had us suppose. Though PostgreSQL has all the time had an ideal group, it has additionally had a refrain of critics. In the meantime, each main cloud supplier has a PostgreSQL database service. Going additional, in 2021 AWS launched Babelfish, an open supply mission that makes it easy to drop in PostgreSQL to functions written for SQL Server.

In like method, there’s a cause that each main cloud supplier affords MongoDB in some kind and that the database has been downloaded extra instances previously 12 months than within the earlier 12 years mixed (for a grand complete of 265 million downloads so far). Each PostgreSQL and MongoDB have made dramatic features in reputation relative to Oracle and SQL Server.

I don’t wish to fanboy this text. However when you’ll indulge me, I’d like to catch you up on a MongoDB you may not know and end with a suggestion that simply would possibly shock you: MongoDB now leads the {industry} in safety, given the discharge of Queryable Encryption.

Atomicity, transactions, and so forth.

I left MongoDB in 2014, proper earlier than the corporate hit overdrive on updates to the core database. The corporate had all the time been a developer darling as a result of its comfort, however round this time MongoDB introduced the acquisition of WiredTiger, which paved the way in which for document-level concurrency management and compression, beginning in MongoDB 3.0 (WiredTiger as an choice) and evolving with MongoDB 3.2 (WiredTiger because the default).

A lot of the work of deeply integrating WiredTiger into MongoDB occurred in 2015, organising a gentle drumbeat of database enhancements for the subsequent few years, with one among my absolute favorites—multi-document ACID transactions—arriving in MongoDB 4.0. As MongoDB cofounder Eliot Horowitz wryly wrote in 2018, “MongoDB drops ACID,” and the MongoDB world was by no means fairly the identical thereafter. In MongoDB 5.0, the corporate launched a versioned API that allowed builders to improve the database with out having to vary their software, time sequence performance, a brand new serverless providing, reside resharding, and extra.

Most not too long ago, at MongoDB World, the corporate introduced plenty of issues to make builders’ lives simpler: Atlas Knowledge API, serverless situations, Atlas CLI, and extra. For me, the 2 most fascinating thematic bulletins have been within the areas of analytics and safety. Oh, and open supply.

Analytics, in MongoDB’s world, are all about enabling builders to construct higher apps, not about information analysts doing offline evaluation. In as we speak’s world, the necessity for a real-time view of the enterprise pushes information evaluation and reporting wants nearer to the functions that generate the info. MongoDB clearly acknowledged this rising want and introduced a number of new methods to make operating analytics in opposition to operational information straightforward, together with a brand new SQL interface, Atlas Analytics Node Tiers, Atlas Knowledge Federation, and Atlas Knowledge Lake.

Even so, it strikes me that the corporate’s key announcement within the space of making use of analytics to real-time information was Columnstore Indexing. Since columnar codecs are perfect for operating analytical workloads, this indexing choice makes it straightforward for builders to maintain paperwork in the suitable mannequin for his or her functions with out shifting the info, and execute performant analytical queries in opposition to that information in actual time for his or her functions.

If all this comes as a shock to these within the “MongoDB is net scale” camp, Queryable Encryption is an excellent greater shocker.

Upping the ante on information safety

Nothing that MongoDB (or any database firm) releases can be of a lot use if it couldn’t match options and performance with safety. MongoDB has lengthy supplied wonderful safety, however the introduction of structured encryption for field-level encryption strikes issues to a different degree. Most databases have discovered safe information at relaxation or in movement however fail to safe information whereas in use when it’s weak to insider entry and lively database breaches. Enter field-level encryption. Area-level encryption protects information in reminiscence and on disk on the server. It’s the best degree of safety for breaches, however it has a draw back: It doesn’t permit for wealthy, expressive querying of encrypted information. Sure, you are able to do precise equality matches, however you need to use deterministic encryption.

Good, however not almost sufficient.

Researchers have been engaged on this downside since 2001, however this week MongoDB introduced the first-ever commercially obtainable, structured encryption mannequin, known as Queryable Encryption. With such structured encryption MongoDB can remodel the encrypted subject in a cryptographically safe approach such that it will probably retailer nameless metadata permitting expressive and environment friendly queries to be carried out. For instance, structured encryption permits a developer to construct a financial institution software that may discover transactions utilizing a variety of dates or greenback quantities for fraud investigation.

That is best-in-industry stuff and doesn’t go away MongoDB’s developer group behind. In brief, MongoDB’s use of Queryable Encryption helps builders hold their deal with constructing participating, data-driven functions whereas assembly the {industry}’s most demanding information privateness challenges. No PhD in cryptography required.

This might be fascinating in and of itself, however MongoDB took an particularly noteworthy angle with Queryable Encryption: It is going to be 100% open. As Porter declared in his keynote, “We will probably be publishing the code, the algorithms, and the mathematics behind it as a result of we consider in white-box safety, not black-box safety.”

This may occasionally come as a shock to those that nonetheless carp on MongoDB’s license change in 2019. (Builders don’t appear to thoughts, given larger priorities.) However it shouldn’t. MongoDB is a contributor to Apache Lucene, releases WiredTiger underneath an open supply license, and in addition affords its Realm cellular information retailer as open supply. It’s straightforward to color corporations in binary, however it’s often incorrect, as on this case.

All of it is a great distance of claiming that maybe it’s time to hit “refresh” in your views on MongoDB. No, it’s not going to be the suitable information platform for all your workloads. Nothing is. However it’s good to make that call based mostly on present actuality, not outdated myths.