TL;DR
NoSQL is a database that can store data in a non-relational way. It is a totally different approach to data storage than SQL databases. You need some time to understant the concept, but it is worth it. Int the public cloud you can use Azure Cosmos DB, from AWS DynamoDB. The storage is fast and scalable, but remember it totally different from SQL databases.
Check:
NoSQL is amazing
Check my favourite cases and build features like:
Link shortening
With Cosmos DB, you can set the TTL (TimeToLive) parameter for how the log record will be stored. You don’t need to maintain the solution and build a scheduled delete service to clean it.
Profile storage
For the CIAM solution, you can store the user profile as a document based on the user identifier, access it with low latency, and, if needed, use geo-replication.
Real-time order processing
For mission-critical applications - real-time response and high availability (99.999%). Also, with Zero ETL opportunity, you can forward the data to Fabric or PowerBI without affecting your request unit (RU) consumption.
AI solutions
With Vector Database and vector search, you can implement persistent memory for AI agents or production-level LLM caching.
From the technical perspective
The data will be stored as JSON, so implementation with Node.js will be straightforward, but no worries—with .Net, it is also easy. You must create a partition key for basic scenarios like link-shortening systems or user profiles. Depending on your needs, the structure for the key will be different—maybe an identifier (GUID), email address, or account number.
Key features
- Global distribution - You can replicate your data to any region in the world.
- TTL - You can set the time to live for the document.
- Consistency levels - You can set the consistency level for the read and write operations.