In the big-data industry, 12-year-old MongoDB has long been a NoSQL flag bearer. At this point in the technology’s evolution, it’s clear that its document database is a credible alternative to relational databases as the platform for many transactional and analytics applications. However, the advantages of document databases for these workloads have not been stark enough to precipitate mass enterprise migrations in their direction.
Be that as it may, MongoDB continues to deepen the value proposition for document databases as a key platform in hybridized big data environments within the private, public, hybrid and multicloud environments. Though MongoDB cannot be regarded as an innovator in the cloud data platform wars, it has demonstrated a solid understanding of what enterprise customers need and, judging by announcements and discussions at their conference, it has made prudent bets in the research and development, partnerships and other investments to address emerging requirements.
Sustaining marketplace momentum
The first thing to keep in mind about MongoDB is its success in building a global customer base and partner ecosystem. In that regard, it continues to rack up impressive numbers, both in customer adoption and in the financial bottom line. The company now has more than 14,200 customers worldwide, up from 5,700 a year ago. Downloads of its flagship open-source document database are now greater than 70 million, which is double the number of downloads it reported a year ago. According to DB Engines, MongoDB is now the fifth most popular database in the world, and owing to availability in the Amazon Web Services, Microsoft Azure and Google Cloud Platform public clouds, MongoDB is now the world’s most widely available cloud database.
MongoDB reported record sales performance of $267 million in the latest fiscal year, representing a compound annual growth rate of 49%. In the latest quarter, it boasted year-over-year revenue growth of 78%. Revenues for its MongoDB Atlas cloud database now amount to more $125 million annually, up from $20 million a year ago.
It has also successfully grown its large accounts, boasting 48 percent compound annual growth rate in customers that spend at least an annualized $100,000 on MongoDB products. Penetrating the small to midsized business market, MongoDB now claims that about 23% of its revenue now comes from self-service online subscriptions, which is up 250% from a year ago.
Deepening the core platform capabilities
Before more than 2,000 attendees this week at its annual developer conference, the company announced MongoDB 4.2, which includes the following significant enhancements:
- Support for distributed database operations: Enterprises are insisting on the ability to leverage their database investments into more distributed global architectures. To address that requirement. MongoDB has evolved the multidocument ACID guarantees that it introduced a year ago, adding the ability to execute transactions on clusters that are deployed in multiple data centers. MongoDB 4.2 uses a consistent syntax and API regardless of whether transactions are executed across documents, collections and databases in a replica set, or across sharded clusters.
- Deepening of multicloud-friendly database deployment features: Cloud-native deployments of databases and their applications are gaining traction in the marketplace. To support those growing requirements, MongoDB 4.2 includes an enhancement to its updated Kubernetes operator, which gives users full management over the document database whether it is deployed on-premises or in private, public and hybrid clouds. Information technology administrators can use MongoDB Enterprise Operator for Kubernetes and MongoDB Ops Manager to automate, manage and replicate data across MongoDB clusters on such popular Kubernetes distributions as Red Hat OpenShift and Pivotal Container Service. It integrates with Kubernetes Open Service Broker, which enables more seamless interface of its Atlas cloud-database APIs into containerized cloud-native environments. And it integrates with Hashicorp’s Terraform infrastructure-as-code environment and Vault key store, thereby enabling secure deployments of database instances on Kubernetes and other platforms in hybrid and multiclouds.
- More stringent database security and compliance: In a world where General Data Protection Regulation and other privacy mandates are bearing down on enterprises, MongoDB has addressed these head-on in its latest announcements. The new MongoDB 4.2 implements field-level encryption, which, executed entirely on the client, enables users to encrypt fields on the server and store them as ciphertext in memory, in system logs or in backups. Field-level encryption is totally separated from the database, transparent to the server and beyond the prying eyes of any IT administrators who have access to the server operating system, database instance, logs and backups. It enables application code to be run unmodified for most database read and write operations that use field-level encryption. It also facilitates “right to be forgotten” compliance by enabling IT administrators to simply destroy a customer’s key, thereby rendering the associated personal data inaccessible. Other new security features in MongoDB 4.2 include support for tighter access controls (multicertificate authorities, forward secrecy TLS, zero downtime certificates, keyfile rotation and other means) and more compliant and efficient auditing (through ISO 27001 certification, lower auditing overhead and other methods).
Expanding the total addressable market
MongoDB’s customers use its core database and other products for a wide range of deployments to support systems of engagement, record, and insight. The company now presents itself as a “full data platform company,” not just a NoSQL database vendor. However, it remains entirely focused on its core NoSQL document database, now in version 4.2, and shows no interest in diversifying into relational, columnar, key-value, graph or any of the other data platform architectures on the market.
In its announcements this week, it introduced new offerings that will help it position MongoDB 4.2 and the Mongo Atlas cloud database as a foundation for enterprise applications in several areas:
- Full-text search applications: In recognition of the fact that its document database often feeds search and knowledge management platforms, MongoDB this week announced the beta of MongoDB Atlas Full-Text Search. This offering embeds the Lucene open-source search engine, giving end users the ability to filter, rank and sort through their data to quickly surface the most relevant results. It provides rich text search capabilities based on Apache Lucene 8 against fully managed MongoDB databases with no additional infrastructure or systems to manage. Once indexes have been created using either the Atlas UI or API, developers can run sophisticated search queries using MongoDB Query Language.
- Mobile data apps: Trying to position itself as a vendor of mobile-data solutions, MongoDB this week unveiled its plans to merge its recently acquired Realm mobile database and synchronization environment with its year-old Stitch serverless platform. Under the MongoDB Realm brand, the forthcoming converged serverless offering will make it easier for developers to work with data all the way from the mobile/web front end to the cloud or on-premises backend. The company is reengineering Realm’s synchronization protocol to connect with the MongoDB Atlas global cloud database as a much more scalable back-end. MongoDB intends to position the reengineered Realm Database as a mainstream platform for mobile developers to build real-time data applications in the browser, and in iOS and Android mobile devices. It will be investing in the core Realm database and intends to deliver the new Realm Sync in 2020. It also intends for these products to serve as the foundation for embedded Internet of Things applications.
- Enterprise data lakes: Positioning its document database as a platform for AI/machine learning app developers, MongoDB this week announced the beta of MongoDB Atlas Data Lake. This new serverless offering supports rich data analytics via the MongoDB Query Language. It supports multiple polymorphic data in multiple schema-free formats at any scale, compressed or uncompressed. It will support a consolidated user interface and billing with on-demand usage-based pricing. For storage, MongoDB Atlas Data Lake allows customers to “bring your own bucket” such as AWS S3, with MongoDB only charging customers for the ability to query the stored data through the Data Lake Service. It allows customers to query data quickly on S3 in any format, including JSON, BSON, CSV, TSV, Parquet and Avro, using the MongoDB Query Language. By bringing the MongoDB Query Language to the MongoDB Atlas Data Lake, this service enables developers to use that language across data on S3, making the querying of massive data sets easier and more cost-effective. The vendor plans availability of MongoDB Atlas Data Lake on Google Cloud Storage and Azure Storage for some point in the indefinite future.
- Streaming data apps: MongoDB announced the beta of an Apache Kafka Connector for robust streaming-data pipelines for microservices and event-driven architectures. The connector enables the MongoDB database to serve as a sink and a source for Kafka streams. It can integrate with change streams and Atlas triggers to create fully reactive, event-driven pipelines.
Though Wikibon lauds MongoDB for pushing the envelope of its own use cases, we have the following concerns surrounding its latest announcements:
- Schemaless document databases exacerbate data silos: MongoDB’s customer base may eventually unravel because of the database consistency issues that are intrinsic to document databases. MongoDB’s dynamically typed schema — also known as “schemaless” architectures — offers developer benefits in easy setup and flexible data modeling. But the downstream cost often rears its ugly head in the fact that, without a schema, responsibility for database consistency shifts to applications. Considering that the agile development practices of the 21st century may run roughshod over database consistency, NoSQL architectures may exacerbate enterprise data siloes. That creates the acute risk of data silos, plus corresponding code duplication, unless strong data governance practices are instituted at the same time that NoSQL databases such as MongoDB are adopted. Wikibon hasn’t seen much movement by NoSQL users in this direction, which means that enterprises may at some point need to migrate away from MongoDB and equivalent NoSQL platforms to data platforms that guarantee relational consistency and referential integrity.
- NoSQL databases have begun to plateau in market acceptance: MongoDB may find itself challenged to grow its customer base outside its well-defined niche. To its credit, it stands apart from the NoSQL vendor field owing to its ability to build a substantial business for its open-source document database. However, the cloud data world has evolved toward hybrid environments in which the various flavors of NoSQL, which started as alternatives to relational, have settled into their own niches in terms of the apps and workloads for which enterprises are deploying them. But relational databases such as Oracle, Microsoft SQL Server and IBM DB2 have evolved over this same period to address the issues that allowed NoSQL alternatives to get a foot in door. And the major public cloud providers have continued to improve their own NoSQL offerings (e.g, AWS DynamoDB) to fend off the competitive threat from MongoDB etc. As we move into the 2020s, it’s clear that MongoDB, Redis, Cassandra, InfluxDB and other NoSQL offerings will fight for diminishing opportunities in enterprise systems of record, engagement, and insight.
- Customers may hold fast to legacy investments in visualization, search, mobility and data lake solutions: MongoDB may be “too little, too late” to many ancillary niches into which it’s growing its solution portfolio. As it delivers its new offerings to market, MongoDB will face formidable competition from many established vendors in each of these niches. In fact, each of MongoDB’s core public-cloud partners — AWS, Microsoft Azure, and Google Cloud Platform — has offerings in each segment, which those partners have every incentive to sell against MongoDB’s 1.0 offerings. For example, AWS has a recently released MongoDB competitor called DocumentDB, The best that MongoDB will be able to do is cross-sell these offerings to existing database customers who need “good enough” features that seamlessly leverage the data they store in its document database. It’s not at all clear that MongoDB has the vision to differentiate these solutions against startups and established providers in the niches in which they’re now competing. For example, MongoDB calls one of its new offerings a “data lake” solution, but it’s not fully baked in the sense that it lacks the necessary data science workbench such as AWS SageMaker to drive development of the AI/machine learning apps intrinsic to the very definition of what a data-lake platform exists to support. Wikibon is concerned that MongoDB doesn’t even have any partnerships, roadmaps or visions in that direction.
Recommendations for users
Many MongoDB customers have built their businesses on the robust transactional and analytic capabilities of its flagship NoSQL database, both on-premises and in the public cloud. Those investments are safe, owing to the company’s strong financials, high-quality engineering, extensive sales and support ecosystem, and keen focus on tooling for operational reliability, security, compliance and optimization.
Customers that have deployed the on-premises enterprise and community editions of MongoDB should consider migrating their licenses to MongoDB running in the AWS, Microsoft or Google public clouds for the potential to improve service levels and cost efficiencies from relying on those providers’ platform-as-a-service offerings. New MongoDB customers should use the self-service pay-as-you-go MongoDB Atlas managed service to get started with this scalable cloud database in an on-demand environment while mitigating the operational and business risks.
For MongoDB customers that want to evolve their investments to support data-driven visualization, full-text search, mobile data apps or enterprise data lakes, bear in mind that the offerings announced this week are either brand-new to market or are simply in beta testing and won’t be generally available until later this year. Just as important, you’ll find many established, mature commercial solutions in the market that support these applications and workloads, and many of them are either from MongoDB partners or can pull data that’s maintained in MongoDB clusters. So far, there is nothing particularly noteworthy about any of the announced solutions in their respective niches.
If you’re seeking out a cloud document database for the development advantages of a multimodel schemaless architecture, you should pay special attention to MongoDB, which is by far the most mature and widely adopted product in this segment. But bear in mind that the range of commercial alternatives from such public cloud providers as AWS and enterprise IT vendors such as IBM Corp. might make it worth your while to include them in your evaluations.