Cloud Equivalents

NoSQL (key-value, document, wide-column, graph)

Amazon DynamoDB vs Azure Cosmos DB for NoSQL

Amazon DynamoDB maps to Azure Cosmos DB for NoSQL.

Both partition data by key, scale throughput elastically, offer serverless and provisioned capacity, and provide change streams and multi-region replication. Their data models, consistency, indexing, and pricing differ, so plan data-model changes rather than a straight port.

Direct match Use it to: Store items by partition key with single-digit-millisecond reads and writes at any scale, without managing servers.

Last verified: · Edit on GitHub · Report an error

Side-by-side comparison

Serverless key-value and document database: AWS vs Azure side by side
AspectAmazon Web ServicesMicrosoft Azure
ServiceAmazon DynamoDBAzure Cosmos DB for NoSQL
LifecycleGenerally availableGenerally available
What it isServerless key-value and document database with single-digit-millisecond performance, on-demand or provisioned capacity, transactions, point-in-time recovery, TTL, global tables, and streams.Globally distributed NoSQL document database with request-unit (RU) throughput (provisioned, autoscale, or serverless), automatic indexing, five consistency levels, multi-region writes (99.999% SLA), and integrated DiskANN vector search.
Pricing modelOn-demand per million read and write request units, or provisioned capacity per hour (with reserved capacity), plus storage per GB-month and optional features (backups, streams, global tables).Provisioned RU/s per hour (standard or autoscale, with reserved capacity), or serverless per million RUs, plus storage per GB-month, multiplied by the number of regions.
DocsAmazon DynamoDB docsAzure Cosmos DB for NoSQL docs
Pricing pageAmazon DynamoDB pricingAzure Cosmos DB for NoSQL pricing
ExamsSAA-C03 DVA-C02 Retiring DEA-C01DP-420 AZ-305

Key differences

  • Capacity units. DynamoDB uses on-demand (per request) or provisioned read and write capacity units. Cosmos DB uses request units (RUs) for all operations, with provisioned, autoscale, or serverless accounts.
  • Indexing and queries. DynamoDB queries by key, with local and global secondary indexes you define. Cosmos DB indexes every property by default and supports a SQL-like query language, including cross-partition queries.
  • Consistency. DynamoDB offers eventually consistent or strongly consistent reads per request. Cosmos DB has five levels (strong, bounded staleness, session, consistent prefix, and eventual), with session as the default.
  • Limits. DynamoDB items are limited to 400 KB, and Cosmos DB items to 2 MB. Both cap throughput per partition, so key design matters on both.
  • Vector search. Cosmos DB for NoSQL has built-in DiskANN vector indexing. DynamoDB typically pairs with OpenSearch or S3 Vectors for vectors.

Pricing summary

Pricing models only. No prices are quoted because they change by region and over time.

AWS
On-demand per million read and write request units, or provisioned capacity per hour (with reserved capacity), plus storage per GB-month and optional features (backups, streams, global tables). Amazon DynamoDB pricing →
Azure
Provisioned RU/s per hour (standard or autoscale, with reserved capacity), or serverless per million RUs, plus storage per GB-month, multiplied by the number of regions. Azure Cosmos DB for NoSQL pricing →

Certification exam mapping

ExamProviderLevelStatus
AWS Certified Solutions Architect – Associate SAA-C03AWSAssociatecurrent
Italian version retires after Dec 31, 2026.
AWS Certified Developer – Associate DVA-C02AWSAssociateretiring
Being updated to DVA-C03. Registration for DVA-C03 opens Oct 27, 2026. Last day to take DVA-C02 is Nov 30, 2026.
AWS Certified Data Engineer – Associate DEA-C01AWSAssociatecurrent
Microsoft Certified – Azure Cosmos DB AI Developer Associate DP-420AzureAssociatecurrent
Renamed from Azure Cosmos DB Developer Specialty and moved to Associate level on Oct 6, 2026.
Microsoft Certified – Azure Solutions Architect Expert AZ-305AzureExpertcurrent
English version updated Apr 17, 2026.

Relevance is based on each exam's published scope, not a line-by-line objective match. Check the official study guide.

FAQ

What is the Azure equivalent of Amazon DynamoDB?

Amazon DynamoDB maps to Azure Cosmos DB for NoSQL. Both partition data by key, scale throughput elastically, offer serverless and provisioned capacity, and provide change streams and multi-region replication. Their data models, consistency, indexing, and pricing differ, so plan data-model changes rather than a straight port.

How does pricing compare between Amazon DynamoDB and Azure Cosmos DB for NoSQL?

AWS: On-demand per million read and write request units, or provisioned capacity per hour (with reserved capacity), plus storage per GB-month and optional features (backups, streams, global tables). Azure: Provisioned RU/s per hour (standard or autoscale, with reserved capacity), or serverless per million RUs, plus storage per GB-month, multiplied by the number of regions. Check the official pricing pages for current rates.

Which certification exams cover this topic?

AWS Certified Solutions Architect – Associate (SAA-C03); AWS Certified Developer – Associate (DVA-C02) [retiring]; AWS Certified Data Engineer – Associate (DEA-C01); Microsoft Certified – Azure Cosmos DB AI Developer Associate (DP-420); Microsoft Certified – Azure Solutions Architect Expert (AZ-305).

Is Cosmos DB the same as DynamoDB?

They solve the same problem and Microsoft maps them, but they aren't API-compatible. Cosmos DB for NoSQL has its own SDKs, query language, and RU-based pricing. Microsoft publishes guidance for migrating DynamoDB apps to Cosmos DB.

Related comparisons

Official sources