Choose a consistency model

A consistency model defines how changes made to a database are propagated and made visible to other clients. In the context of designing and implementing native applications using Microsoft Azure Cosmos DB, you have the flexibility to choose from multiple consistency models to suit your application’s requirements. This article will explore the available consistency models and provide insights into their usage.

Multi-Master Replication in Azure Cosmos DB

Before delving into the consistency models, it’s essential to understand the multi-master replication aspect of Azure Cosmos DB. Azure Cosmos DB is a globally distributed NoSQL database that allows you to replicate your data across multiple regions worldwide. This replication enables local reads and writes, ensuring low-latency access to the data and high availability.

Available Consistency Models

1. Strong Consistency

   • In this model, the database guarantees linearizability, which means that a read operation always returns the most recent write.
   • Strong consistency ensures that all clients see the same order of operations, providing the highest level of data consistency.
   • However, it might impact performance and increase latency due to synchronization delays across regions.
   • To use strong consistency, you can set the “ConsistencyLevel” property to “ConsistencyLevel.Strong” in the Cosmos DB SDK.

2. Bounded Staleness Consistency

   • Bounded staleness consistency allows you to configure the acceptable lag (in terms of versions or time) for read operations.
   • It provides a compromise between strong consistency and higher performance by allowing some lag in data synchronization.
   • You can specify the maximum permissible staleness window to read data, balancing consistency and performance requirements.
   • By setting the “ConsistencyLevel” property to “ConsistencyLevel.BoundedStaleness,” you can configure the staleness window and the maximum lag values.

3. Session Consistency

   • Session consistency guarantees linearizability for each individual session, meaning that all operations within a session are seen in the order they were performed.
   • It ensures strong consistency for clients within their own session, providing a consistent view of the data.
   • However, different sessions might observe operations in a different order, which allows for higher performance and reduced latency compared to strong consistency.
   • To set session consistency, you can set the “ConsistencyLevel” property to “ConsistencyLevel.Session”.

4. Consistent Prefix Consistency

   • The consistent prefix consistency model guarantees that clients always observe a prefix of the operations in the order they were performed.
   • It offers monotonic reads, meaning that once a client sees a particular state of data, it will not see older states in subsequent reads.
   • Consistent prefix consistency is useful when strong consistency is not necessary, and higher read performance is desired.
   • You can set the “ConsistencyLevel” property to “ConsistencyLevel.ConsistentPrefix” to use this model.

5. Eventual Consistency

   • Eventual consistency allows for the highest level of availability and lowest latency at the cost of weaker consistency guarantees.
   • In this model, there is no ordering guarantee, and different clients may observe data changes in different orders.
   • It is well-suited for scenarios where immediate strong consistency is not a requirement, such as handling user-generated content or analytics data.
   • To use eventual consistency, set the “ConsistencyLevel” property to “ConsistencyLevel.Eventual”.

Here’s an example of how you can choose a consistency model when working with Azure Cosmos DB in a .NET application:

using Microsoft.Azure.Cosmos;
using Newtonsoft.Json;

[JsonProperty("name")]
public string Name { get; set; }
}


In the code snippet above, the consistency level is explicitly set to “ConsistencyLevel.Strong” using the CosmosClientOptions. You can change this value to use other consistency models according to your application’s requirements.

Remember, the choice of consistency model depends on the trade-offs between consistency, availability, and performance that your application demands. Azure Cosmos DB provides the flexibility to choose an appropriate consistency model to achieve the desired balance.

In conclusion, designing and implementing native applications using Microsoft Azure Cosmos DB requires careful consideration of the consistency model. By understanding the available options, you can make informed decisions to provide the desired level of consistency, performance, and availability for your application.