| Data Model | Primarily Key-Value and Document | Multi-model: Document, Key-Value, Wide-Column (Cassandra API), Graph (Gremlin API), Table (Table API) | Document-oriented | Wide-column (Column-family) |
| Scalability | Highly scalable, automatic partitioning (Partitioning) | Massively scalable, automatic and instant scalability with global distribution options (Partitioning) | Highly scalable, automatic scaling (Add Data – scaling is implicit) | Massively scalable, horizontal scaling via nodes (Scalability) |
| Global Distribution | Multi-region, multi-active with Global Tables | Built-in global distribution with multi-region writes (multi-master) available (Global Data Distribution) | Multi-region with strong consistency options (Locations) | Global replication available (Replication Overview) |
| Consistency | Tunable (Read Consistency): Eventual, Strong | Five well-defined models (Consistency Levels): Strong, Bounded Staleness, Session, Consistent Prefix, Eventual | Tunable (Transactions – implies strong), Eventual | Strong consistency per row (Consistency) |
| Querying | Key-based lookups, Scan, Query API with limited filtering | SQL-like (SQL Query) for document model, API-specific for others | Rich querying with indexing, including compound indexes | Key-based lookups, range scans, filtering on columns |
| Transactions | ACID transactions (multi-item) (Transactions) | ACID transactions across documents within a partition (Transactions) | ACID transactions (multi-document) (Transactions) | Transactions within a single row (Transactions) |
| Serverless | Fully serverless, auto-scaling to zero (On-Demand Pricing) | Fully managed with serverless options and provisioned throughput with auto-scaling | Fully serverless, pay-per-use (Firebase Pricing) | Not fully serverless, billed by node hours (Pricing) |
| APIs/Ecosystem | AWS ecosystem integration (AWS) | Broad API support, strong Azure integration (Azure), integration with Microsoft Fabric | GCP ecosystem integration (Google Cloud), Firebase integration | GCP ecosystem integration (Google Cloud), HBase compatibility |
| Pricing | Provisioned (Provisioned Pricing) and On-demand based on RCUs/WCUs and storage | Provisioned Throughput (Provisioned Pricing) and Serverless based on request units, storage, and bandwidth | Pay-per-read, pay-per-write, pay-per-storage (Firebase Pricing) | Billed by node hours, storage, and network egress (Pricing) |
| Indexing | Primary key (partition and sort key), Global Secondary Indexes (GSIs), Local Secondary Indexes (LSIs) | Automatic indexing of all attributes by default, with options to customize (Index Policy) | Automatic indexing with options for composite and single-field indexes | Row key, column families, and qualifiers are indexed (Schema Design – key design is crucial for querying) |
| Managed Services | Fully managed (Features) | Fully managed (Overview), handles patching, upgrades, backups, global replication | Fully managed (Overview), handles patching, upgrades, backups, multi-region | Fully managed (Overview), handles patching, upgrades, backups, replication |
