- \n
- Real-time performance<\/b>: Memory-first architecture with built-in caching delivers low-latency read and write operations, see our benchmark reports:\n
- \n
- 3rd party benchmark – Altoros<\/a><\/li>\n
- 3rd party benchmark – BenchAnt<\/a><\/li>\n<\/ul>\n<\/li>\n
- Powerful query language<\/b>: SQL++ with supporting JOINs, aggregations, and complex queries<\/li>\n
- AI Services<\/b>: Design to accelerate AI development and integration within your apps<\/li>\n
- Real-time analytics<\/b>: Massively parallel processing with a JSON-native columnar engine<\/li>\n
- Proven mobile solution<\/b>: Used by many of the world\u2019s leading companies\n
- \n
- MongoDB has recently deprecated their Realm product<\/b> line, ending support in September of 2025<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
Customer examples who chose Couchbase over MongoDB:<\/p>\n
- \n
- \n
- \n
- Small and medium-sized businesses\n
- \n
- Egaming customer<\/a><\/li>\n
- Jinmu<\/a> – IT Services, AI assistant<\/li>\n
- Microlise<\/a> – Fleet and logistics management<\/li>\n<\/ul>\n<\/li>\n
- Enterprises\n
- \n
- FICO<\/a> – Real-time fraud detection<\/li>\n
- Cisco<\/a> – On-demand TV service<\/li>\n
- Equifax<\/a> – Credit lending platform<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
How to move from MongoDB to Couchbase<\/h3>\n
The following is a strategy<\/b> to aid your migration, starting with the basics and then following up with the actual migration process and post-migration steps.<\/p>\n
The basics<\/h2>\n
Understand the key differences<\/strong><\/p>\n
- \n
- \n
- \n
- Data Model<\/b>: Both are document-based. Couchbase uses JSON documents<\/b>; MongoDB uses BSON (which can be exported to JSON).<\/li>\n
- Query Language<\/b>: MongoDB uses MQL (MongoDB Query Language)<\/b>, while Couchbase uses SQL++, <\/b>which extends industry standard SQL to work with JSON<\/li>\n
- Architecture<\/b>: Couchbase has memory-first architecture<\/b>, built-in distributed caching<\/b>, and stronger support for scalability<\/b><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
Define your migration goals<\/strong><\/p>\n
- \n
- \n
- \n
- Are you migrating everything<\/b> or only a subset<\/b> of data?<\/li>\n
- How will you minimize downtime<\/b> during migration?<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
\nThe migration (data and application changes)<\/h2>\n
Choose a data migration approach<\/h4>\n
There are two main strategies:<\/p>\n
- \n
- \n
- \n
- Direct data migration<\/b>, which involves copying a snapshot of data from MongoDB and importing into Couchbase<\/li>\n
- Incremental migration<\/b>, which involves live synchronization of data from MongoDB into Couchbase<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
Direct data migration (full dump & load)<\/h5>\n
Recommended if:<\/p>\n
- \n
- \n
- \n
- You have a manageable amount of data<\/b><\/li>\n
- Downtime is acceptable<\/b> during migration<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
A manageable amount of data for a one-time dump and load from MongoDB to Couchbase depends on multiple factors like hardware specs, network speed, and indexing needs. However, here are some general guidelines based on the number of documents:<\/p>\n
- \n
- \n
- \n
- Under 10 million documents<\/b> \u2192 Generally straightforward<\/li>\n
- 10M – 100M documents<\/b> \u2192 Requires tuning batch sizes, indexing strategies (consider post-loading at least some indexes)<\/li>\n
- 100M – 500M documents<\/b> \u2192 May need parallelized loading, post-loading indexes<\/li>\n
- 500M+ documents<\/b> \u2192 Likely better suited for staged migration or streaming approach<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
Steps<\/b><\/p>\n
You can use cbmigrate<\/a>, a tool that’s part of the Couchbase CLI collection<\/a>. However, it will likely be easier to use the Couchbase IDE plugins for VSCode or JetBrains, which provide development tools, as well as a GUI wrapper for cbmigrate<\/i>. Here are the steps for using VSCode:<\/p>\n
- \n
- \n
- \n
- Install VSCode Couchbase extension<\/a><\/li>\n
- Connect to the Couchbase cluster that you wish to migrate MongoDB data\/indexes into. If you don’t have a cluster setup already, you can create a free cluster on Couchbase Capella<\/a>.
\n![\"Migrate](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image2.png\")
<\/li>\n- Right-click on the cluster connection, select Tools<\/b> > MongoDB Migrate
\n![\"MongoDB](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image5.png\")
<\/b><\/li>\n- Enter your MongoDB connection information, press Connect<\/b>, and select the database\/collections you wish to migrate. Then, select the target location in Couchbase (bucket + scope). You can also migrate indexes, but this is optional.
\n![\"Data](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image1.png\")
<\/li>\n- Click Migrate<\/b> to continue. In the terminal output, you will notice that cbmigrate<\/b> is being used.
\n![\"\"](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image4.png\")
<\/li>\n- You can use the Couchbase plugin to continue to interact with the migrated data. For instance, to browse collections, documents, and indexes.
\n![\"\"](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image3.png\")
<\/li>\n<\/ol>\n<\/li>\n<\/ol>\nIncremental migration (live sync)<\/h5>\n
Recommended if:<\/p>\n
- \n
- \n
- \n
- Your dataset is large<\/b><\/li>\n
- Zero downtime is required<\/b><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
There are multiple methods and tools for streaming data from MongoDB to Couchbase. A common tool that your organization may already be using is Kafka (and\/or Confluent). Couchbase provides an officially supported Kafka connector<\/a>, and MongoDB provides an officially supported connector<\/a> as well.<\/p>\n
Steps<\/b><\/p>\n
- \n
- \n
- \n
- You will need a Kafka (with Kafka Connect, Zookeeper, etc) or Confluent instance already installed and setup<\/li>\n
- Install the MongoDB source<\/b> connector (place the JAR file in the plugin path, or enable the MongoDB connector in Confluent<\/a>)<\/li>\n
- Setup connector configuration<\/a>. Consult the documentation for all the source <\/b>options<\/a>. It will look similar to:\n
\"config\": {\r\n \"connector.class\": \"com.mongodb.kafka.connect.MongoSourceConnector\",\r\n \"connection.uri\": \"mongodb:\/\/mongoconnectionstring:27017\",\r\n \"database\": \"yourDatabase\",\r\n \"collection\": \"yourCollection\",\r\n \"topic.prefix\": \"mongo\",\r\n \"startup.mode\": \"copy_existing\",\r\n \"poll.max.batch.size\": 1000,\r\n \"poll.await.time.ms\": 500\r\n}<\/pre>\nNote<\/b>: if you want to migrate all collections, omit the collection property from configuration.<\/li>\n
- Install the Couchbase sink <\/b>connector (place the JAR file in the plugin path, or enable the Couchbase connector in Confluent<\/a>)<\/li>\n
- Setup connector configuration<\/a>. Consult the documentation for all the sink <\/b>options<\/a>. It will look similar to:\n
\"config\": {\r\n \"connector.class\": \"com.couchbase.connect.kafka.CouchbaseSinkConnector\",\r\n \"tasks.max\": \"1\",\r\n \"connection.cluster_address\": \"couchbase:\/\/cbconnectionstring\",\r\n \"connection.bucket\": \"yourBucket\",\r\n \"connection.username\": \"yourUsername\",\r\n \"connection.password\": \"yourPassword\",\r\n \"topics\": \"mongo.yourDatabase.*\",\r\n \"couchbase.topic.to.collection\":\r\n \"mongo.yourDatabase.users=default_scope.users, ... etc ...\",\r\n \"couchbase.document.id\": \"${\/id}\",\r\n \"couchbase.persist.to\": \"NONE\",\r\n \"couchbase.replicate.to\": \"NONE\",\r\n \"couchbase.retry.max.duration\": \"5m\"\r\n}\r\n<\/pre>\n<\/li>\n<\/ol>\n<\/li>\n<\/ol>\nNote<\/b>: This configuration will subscribe to all topics (which each correspond to a collection in Mongo and a collection in Couchbase). Make sure the collections already exist in Couchbase.<\/p>\n
Once these connectors have been set up, data will start to flow from Mongo to Couchbase. Any changes to Mongo will automatically make their way to Couchbase as well. This allows you to run both concurrently, keeping Couchbase up-to-date to the point you are ready to switch over the application.<\/p>\n
\nUpdate your application code<\/h2>\n
Whether you use the incremental or direct migration approach, you will likely need to update your application code.<\/p>\n
This may include:<\/p>\n
- \n
- \n
- \n
- Modify queries to use SQL++<\/b> instead of MQL<\/b>.<\/li>\n
- Use Couchbase SDKs<\/b> for Python, Java, Node.js, etc.<\/li>\n
- Use Couchbase framework integrations for Spring, EF Core, Node.js, etc.<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
There is no “easy button” for this portion of the migration. The length of the process will depend on how the code is architected, which framework(s) the application is using, and how many queries need to be converted from Mongo’s MQL language to Couchbase’s SQL++.<\/p>\n
Here are some general tips and tools:<\/p>\n
Use the Couchbase SDKs<\/h3>\n
Couchbase has SDKs<\/a> available for 11 languages, including .NET, Go, Java, Node.js, and more. You will need to add the appropriate SDK to your project. There are quick starts available for each SDK. For instance, with .NET, you can use NuGet<\/a>; for Java, you can use Maven<\/a>; for Node.js, you can use npm<\/a>; etc.<\/p>\n
The table below shows examples in Node.js. Couchbase’s approach to SDKs is to be as idiomatic to the language as possible, but basic operations will look similar in all SDKs.<\/p>\n
- Use Couchbase SDKs<\/b> for Python, Java, Node.js, etc.<\/li>\n
- Modify queries to use SQL++<\/b> instead of MQL<\/b>.<\/li>\n
- Setup connector configuration<\/a>. Consult the documentation for all the source <\/b>options<\/a>. It will look similar to:\n
- Zero downtime is required<\/b><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
- Your dataset is large<\/b><\/li>\n
- Connect to the Couchbase cluster that you wish to migrate MongoDB data\/indexes into. If you don’t have a cluster setup already, you can create a free cluster on Couchbase Capella<\/a>.
- Install VSCode Couchbase extension<\/a><\/li>\n
- 10M – 100M documents<\/b> \u2192 Requires tuning batch sizes, indexing strategies (consider post-loading at least some indexes)<\/li>\n
- Under 10 million documents<\/b> \u2192 Generally straightforward<\/li>\n
- Downtime is acceptable<\/b> during migration<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
- You have a manageable amount of data<\/b><\/li>\n
- Incremental migration<\/b>, which involves live synchronization of data from MongoDB into Couchbase<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
- Direct data migration<\/b>, which involves copying a snapshot of data from MongoDB and importing into Couchbase<\/li>\n
- How will you minimize downtime<\/b> during migration?<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
- Are you migrating everything<\/b> or only a subset<\/b> of data?<\/li>\n
- Query Language<\/b>: MongoDB uses MQL (MongoDB Query Language)<\/b>, while Couchbase uses SQL++, <\/b>which extends industry standard SQL to work with JSON<\/li>\n
- Data Model<\/b>: Both are document-based. Couchbase uses JSON documents<\/b>; MongoDB uses BSON (which can be exported to JSON).<\/li>\n
- Cisco<\/a> – On-demand TV service<\/li>\n
- Jinmu<\/a> – IT Services, AI assistant<\/li>\n
- Egaming customer<\/a><\/li>\n
- Small and medium-sized businesses\n
- \n
- 3rd party benchmark – BenchAnt<\/a><\/li>\n<\/ul>\n<\/li>\n
- 3rd party benchmark – Altoros<\/a><\/li>\n
![\"AI](\"https:\/\/www.couchbase.com\/blog\/wp-content\/uploads\/sites\/1\/2025\/04\/image6.png\")
<\/p>\n