Overview

Amazon Keyspaces is a managed database service for applications built on Apache Cassandra’s data model and query interface. It is best suited to teams moving Cassandra workloads or building services that need Cassandra compatibility without taking on infrastructure management. The trade-off is a choice between traffic-based and provisioned capacity charges, with only a time-limited free tier for getting started.

Key features

Keyspaces supports the Cassandra Query Language API, existing application code, drivers and developer tools. That gives Cassandra teams a route to managed operations without replacing their application interface. It is less compelling if Cassandra compatibility is not a requirement, since that is the service’s defining advantage.

Tables can run in on-demand mode, which adjusts capacity to application traffic and charges for reads and writes performed, or provisioned mode, where teams specify expected read and write capacity. Provisioned capacity can use auto scaling. On-demand suits workloads with changing demand when teams want to avoid setting capacity in advance; provisioned mode offers more direct control over capacity settings, but charges for provisioned capacity as well as storage and other applicable usage.

AWS describes single-digit-millisecond read and write performance, and tables have no limit on size or row count. Individual rows are limited to 1 MB, so applications with larger records must account for that constraint. Multi-Region deployments offer up to 99.999% availability, while table data is replicated three times across multiple Availability Zones. These capabilities suit latency-sensitive or geographically distributed services, though the stated availability figure is for Multi-Region deployments.

For operations and data movement, Keyspaces integrates with CloudWatch metrics and can stream changes to Amazon OpenSearch Service. Keyspaces Streams exposes change data through its Streams API or the Kinesis Client Library. Point-in-time recovery maintains continuous backups and can restore a table to any second in the preceding 35 days, a useful recovery window for accidental changes or deletions.

Customer data is encrypted at rest by default, using AWS-owned or customer-managed keys through AWS Key Management Service. Identity and access controls integrate with AWS IAM, and connections can use TLS and AWS PrivateLink to an Amazon VPC. Managed Time to Live settings expire rows or attributes and remove expired records without requiring users to manage tombstones or compaction.

Pricing

On-demand capacity and provisioned capacity have custom pricing rather than a stated monthly price. On-demand billing is based on reads and writes performed, plus storage and other applicable usage. Provisioned billing is based on the capacity specified, storage and other applicable usage; auto scaling can adjust that capacity. Neither mode has a per-month price stated, so the right choice depends on workload demand and the need to control throughput settings.

The AWS Free Tier costs 0.00 USD per free for the first three months after creating the first Keyspaces resource, and is limited to one free tier per payer account. It includes 30 million write request units, 30 million read request units and 1 GB of storage. Those caps make it useful for a small evaluation or low-volume start, not an enduring free deployment; usage beyond the included allowance may incur charges.

Database Savings Plans have custom pricing, require a usage commitment over a 1-year term and state savings of up to 18% by AWS. This may suit workloads with predictable committed usage, but is a less flexible fit when demand or ongoing use is uncertain.

Platforms

Keyspaces is available through API and web interfaces. It is a managed service, with CQL as its query interface; multi-region clusters, tunable consistency controls, transaction support and horizontal scaling are supported.

Who it's for

Keyspaces fits teams migrating Cassandra workloads, or building low-latency applications, IoT device data stores, game-player profiles or time-series systems that benefit from Cassandra compatibility. Its managed, automatically scaling model also suits teams that want to avoid infrastructure management. Teams that need to run and control their own database deployment, or cannot accept capacity-based usage charges and a short free tier, should consider other options.

Pros and cons

Pros

Cassandra compatibility: Existing application code, drivers and developer tools can continue using the Cassandra Query Language API, easing migration.
Managed scaling and regional resilience: On-demand capacity adjusts to traffic, while Multi-Region deployments offer up to 99.999% availability.
Useful recovery and data controls: Point-in-time restore reaches any second within 35 days, and managed TTL avoids user-managed tombstones and compaction.
Flexible capacity choices: Teams can choose traffic-based on-demand billing or specify provisioned capacity and enable auto scaling.

Cons

Cost forecasting requires workload estimates: Both capacity modes have custom pricing, and charges also include storage and other applicable usage.
The free tier is short and capped: It lasts three months, allows one per payer account, and includes only 1 GB of storage alongside request-unit caps.
Rows have a firm size ceiling: The maximum row size is 1 MB, which rules out larger individual records.

Alternatives

Apache Cassandra is a free, self-managed open-source distributed database for teams that prefer to operate Cassandra themselves rather than use Keyspaces’ managed service.

Google Cloud Bigtable is a paid alternative with a free trial and Enterprise and Enterprise Plus plans at 0.65 USD per month and 0.85 USD per month, respectively; consider it when comparing those Bigtable editions with Keyspaces’ Cassandra-compatible approach.

Apache HBase is a free, open-source distributed database whose deployments operators configure and manage, making it an option for teams seeking a self-managed alternative.

SWC-DB is a free, self-hosted option; commercial-purpose licensing inquiries apply to uses limited by GPLv3.

Apache Accumulo is a free, open-source sorted distributed key/value store under the Apache License 2.0.

HPE Ezmeral Data Fabric is a paid option with a free trial; its software pricing varies by reseller, and the Canada store shows “Starting at C$500.00” for offerings with one- to five-year terms.

Baidu Tera is another free, self-hosted alternative.

Cloudera Data Lake Service is a paid alternative that directs buyers to contact sales for pricing.

Browse more options in Wide-Column Databases or Database Software.

Verdict

Choose Amazon Keyspaces if your application depends on Cassandra interfaces and you want a managed service with automatic scaling, recovery controls and multi-region availability. Its strongest reason to buy is retaining Cassandra compatibility while offloading infrastructure work. Look elsewhere if self-management is important, your records exceed 1 MB, or you need a predictable price before estimating real usage.

	Amazon Keyspaces	MarkLogic
Free plan	Yes	Yes
Free trial	No	Yes
Paid from	Free	Free
Open source	No	No
Platforms	api, Web	api, Linux, self-hosted, Web, Windows
Free plan	No	No
SQL support	—	Yes
Vector search	—	Yes

Amazon Keyspaces vs MarkLogic

Amazon Keyspaces

MarkLogic