Azure Cosmos DB for MongoDB

I am Suniti ( linkedin ), a Microsoft Learn student ambassador. In this blog, we’ll take a closer look at the key features and benefits of Azure Cosmos DB for MongoDB and guide developers in choosing the right architecture to scale modern applications.

Let’s say your app is a social media platform for artists where users can upload and share their work, create posts, and interact through comments and likes. The app quickly became popular, especially after several famous artists started using it and talking about it online. Suddenly, thousands of new users are signing up every hour, and your existing MongoDB infrastructure can’t keep up with the massive increase in traffic. With Azure Cosmos DB for MongoDB, you benefit from a highly scalable, globally available, and cost-effective solution while still using familiar MongoDB tools, so you can focus on improving your apps instead of worrying about infrastructure limitations.

There are now two architecture options for scaling MongoDB apps using Azure Cosmos DB for MongoDB: vCore architecture and Request Unit (RU) architecture.

1. vCore architecture:

This option provides a dedicated instance for your MongoDB app, providing friendly scaling and seamless integration with Azure services for users with existing MongoDB knowledge.

• Seamless AI integration: Let’s say your platform introduces an AI feature that recommends artwork based on user preferences. A unified vector database allows you to store both transactional data (user uploads, comments) and vector data (AI-generated recommendations) in the same database. This simplifies AI integration by eliminating the complexity and cost of data transfer between services.
• Efficient text search: Artists often search for specific keywords, such as ‘abstract’, to find similar works. Text indexes allow you to enhance these searches across MongoDB collections, ensuring fast, relevant results without additional complexity.
• Scaling vertically without shard keys: As your app grows and your data size increases, you can scale vertically without worrying about defining partitioning keys, simplifying development and saving time. For example, as users upload larger and higher-resolution artwork, the database can easily grow without any additional management.
• Cost-effectiveness through friendly pricing: You pay based on compute (vCores and RAM) and storage (disk). So, if you’re already using MongoDB, you don’t need to adapt to the new pricing model. Flat pricing makes it easier to manage costs as your platform grows.
• Backup and Restore: with 35 days of backup and point-in-time restore (PITR); If something goes wrong, you’re protected from data loss. For example, if a user accidentally deletes their profile, they can easily restore that data from a backup.

2. Request Unit (RU) architecture:

This architecture is ideal for cloud-native apps and provides a flexible and dynamic scaling approach.

• Instant scalability: As influencers drive traffic surges on your platform. auto resize The feature dynamically adjusts capacity to handle the influx. Unlike other services that can take hours to scale, Azure Cosmos DB can handle increased traffic immediately, ensuring your users don’t experience any delays.
• Automatic sharding: As the platform grows and the database expands, Automatic and transparent With sharding, scaling happens behind the scenes. This means you don’t have to manually manage or configure shards as your app grows to horizontally handle millions of users.
• High availability and global access: with 5 9-valents (99.999%) It provides availability and an active-active database, so your app will continue to run even if part of your infrastructure fails. For example, when users in multiple regions upload artwork simultaneously, Azure Cosmos DB ensures that there is no single point of failure and that your data is always accessible. MongoDB global clusters only support active-passive deployment for writes to the same data.
• Real-time analytics (HTAP): The app’s dashboard tracks user engagement and runs real-time analytics to try to determine what types of artwork are getting the most attention. with Azure Synapse Link, You can do this without affecting platform performance, allowing you to generate instant reports on user behavior.
• Serverless deployment: When traffic is low, such as during off-peak hours, serverless capacity mode means you only pay for work performed. If no one uploads their work at 3am, it’s very cost-effective because you’re not paying for unused resources.

Which architecture should I choose? 🤔

Choose vCore-based if:

• you are Migrate existing MongoDB workloads Or building a new MongoDB application.
• Workloads include long-running queries, complex aggregation pipelines, distributed transactions, or joins.
• Your application requires: 99.995% availability.
• Native support is needed to store and retrieve vector embeddings.

Choose RU-based if:

• You’re developing something new. Cloud-based MongoDB app Or refactor existing apps for a cloud-native environment.
• The workload is focused on point reads (getting a single item by _id and partition key) and has fewer long-running queries or complex aggregation operations.
• you need Unlimited horizontal scalabilityInstant scaling and precise throughput control.
• Your applications demand industry-leading performance. 99.999% availability.

Both options ensure your app can handle increased traffic, provide high availability, and optimize costs based on your specific needs.

Why Cosmos DB?

Azure Cosmos DB and MongoDB It’s both NoSQL A highly available, scalable, and globally distributed database. However, they have different strengths and weaknesses. Here we will discuss why you should move to Cosmos DB for MongoDB.

• Best uptime SLA For MongoDB provided by Cosmos DB
• Instant scale up/down No warm-up period
• Basic integration Azure Synapse Link for analysis
• Supported by Cosmos DB Multiple data models within a single database
• Scalability of Azure Cosmos DB for MongoDB –
  • horizontal expansion – Data and tasks are partitioned across groups of machines. Calls the process of splitting data. Sharding. There are no limits, you can store your data across multiple servers and data centers, and Cosmos DB for MongoDB manages it for you.
  • vertical expansion – As you upgrade to bigger and bigger machines, you will hit your limits.
• Integration with Azure : Azure Cosmos DB for MongoDB is integrated with the Azure ecosystem, allowing developers to use existing MongoDB tools and applications.

conclusion

Azure Cosmos DB is a fully managed database solution designed for modern app development, supporting NoSQL, relational, and vector data models. It delivers single-digit millisecond response times, automatic, on-the-fly scalability, and ensures performance at any scale. It especially powers ChatGPT. Dynamic scalability With high reliability and minimal maintenance

One of the key benefits of Azure Cosmos DB for MongoDB is its comprehensive SLA that covers the entire stack, including the database and underlying infrastructure. This is different from third-party MongoDB services such as MongoDB Atlas, which only deal with the database and exclude any services, hardware or software provided by the cloud platform.

Resource Reference –

1. Introduction/Overview – Azure Cosmos DB for MongoDB | microsoft run
2. MongoDB app has been rebuilt | microsoft run
3. Introducing Azure Cosmos DB for MongoDB vCore: Now generally available! – Azure Cosmos DB Blog
4. Introduction/Overview – Azure Cosmos DB for MongoDB (vCore) | microsoft run
5. Choosing between RU-based and vCore-based models – Azure Cosmos DB for MongoDB | microsoft run
6. Create a free Azure account or pay as you go | Microsoft Azure
7. Use Azure Cosmos DB for free.
8. Analytics with Azure Synapse Link – Azure Cosmos DB | microsoft run