4 min read

Azure Cosmos DB Decoded: Essential Concepts Unveiled

Let's dive into Cosmos DB and understand its features and some of its high-level concepts.
Azure Cosmos DB Decoded: Essential Concepts Unveiled
Photo by NASA / Unsplash


Even today, many organizations use and prefer RDBMS such as MS SQL, MySQL, Oracle, PostgreSQL, etc.

However, there's another approach to data management, commonly known as NoSQL, which developers can't ignore because of this rise in popularity and usage of different companies, from startups to enterprises.  

This article will explore Azure Cosmos Db at least at an introduction level. 

What is Cosmos DB? 

  • It is a globally distributed, horizontal scale and multi-model database
  • It is a No-SQL database to store non-structured data. 
  • It is highly performant and supports automatic scaling, high availability, and low latency. 

Multi-Model Database Support

Multi-model support is one of the key features of Azure Cosmos DB. As a result, it allows developers to store and query data using multi-data models.

In other words, developers can choose the best data model for their application without worrying about whether Azure Cosmos DB supports it.


Stores a collection of key-value pairs.

Each key is distinct and corresponds to a single value.

Key-value data is ideal for applications that store and retrieve data based on the key.

Column Family | Columnar

Columnar uses a collection of columns to store data. 

Each column contains a single data type, such as strings, numbers, or booleans.

Columnar data is well-suited for applications that store and query large, frequently filtered or aggregated datasets.


Document data stores data as a collection of documents. 

Each document is a JSON-like object that can contain a variety of fields. 

It is well-suited for applications that store and retrieve complex data structures.


A graph is a collection of data-storage nodes and edges.

Nodes represent people, places, and things.

Edges represent the connections between nodes.

Graph data is ideal for modeling and querying complicated interactions between items.

What is Global Distribution? 

Global distribution refers to the ability to deploy and manage resources, applications, and data across multiple geographic regions. 

Moreover, global distribution is comparable to replication for relational databases; the difference is that Azure handles everything, and you don't need complex configurations at the database or application levels.

Azure does provide a global network of data centers, and users can deploy resources in multiple geographic regions. As a result, it helps reduce latency, enhance performance, improve availability, and ensure data resiliency. 

Therefore, global distribution within the context of Azure Cosmos Db means that you can deploy your database resources in multiple Azure regions to bring the closure of your services to end-users.

What is Multi-Region Writes?

This "Multi-region Writes" allows organizations to write data to multiple Azure regions simultaneously, improving the application's availability and responsiveness. 

Moreover, Azure automatically handles the distribution and replication of data across specified regions.

What will happen if multi-Region is Enabled? 

Write Replication 

Any data changes are automatically replicated to all the regions. 

In other words, the writes are not limited to a single primary region

and can be distributed across multiple regions in near real-time.

Low-Latency Writes

Applications can write to the nearest available region, reducing write latency. 

It is beneficial for globally distributed applications where it is crucial to minimize the time it takes to write data.

Automatic Failover

Azure Cosmos DB automatically fails to reach a healthy region in a regional outage or disruption. 

Despite regional failures, your application can continue to write and read data.

Storage Redundancy Types

Local Redundancy

It refers to replicating data within the same geographic region to ensure high availability and fault tolerance. In other words, local redundancy focuses on redundancy within a specific region.

Moreover, this redundancy can help protect against hardware failures, network issues, or other localized disruptions.

Geo Redundancy

It refers to replicating data across multiple geographic regions.

It involves redundant copies of data stored in different regions, so if one region experiences an outage or disruption, the data can still be accessed and served from another region.

Moreover, it is more about implementing high availability and disaster recovery strategies.

Cosmos DB Features to Highlight

Globally distributed

Achieve unparalleled data resilience by strategically selecting multiple geolocations for seamless data replication, fostering global accessibility and redundancy.

Low latency

Ensure an exceptional user experience with minimal delay by prioritizing low latency for both data read and write operations, optimizing the responsiveness of your system.

High availability

Maintain data availability by implementing a robust failover concept, guaranteeing a consistent and reliable data access experience, even in the face of potential disruptions.


Enhance data retrieval efficiency with intelligent auto-indexing mechanisms, minimizing latency in read operations. Empower developers with customizable indexing options for fine-tuned performance.

High consistency

Tailor the level of data consistency to your specific needs with five distinct consistency choices, ensuring a flexible and adaptive approach to maintaining the integrity of your data.

Always-on system

Safeguard against downtime by perpetually maintaining system availability through strategically replicating data across multiple geographically dispersed data centers.


Experience economic advantages by choosing Azure Cosmos DB, which offers a cost-effective alternative to AWS DynamoDB, making it a prudent choice for budget-conscious enterprises.


As we have discovered, Azure Cosmos DB is globally distributed, high-performance, and provides multi-model database service.

Moreover, the multi-model database service serves developers without worrying whether Azure Cosmos DB supports it.

Furthermore, we have discussed the multi-region and types of redundancy, which will show why Azure Cosmos DB as a NoSQL is a horizontal-scale database.

I hope you have enjoyed this article, as I wanted it while writing. Till next time, happy cloud computing and programming!