Relational Databases

Amazon RDS vs Cloud SQL

Compare Amazon RDS and Cloud SQL side-by-side. See feature differences and pricing in the Relational Databases category.

Amazon RDS

by Amazon

No reviews yet

Pricing: Usage-based, per instance-hour. MySQL or PostgreSQL on db.t4g.micro in us-east-1: $0.016/hr On-Demand, $0.0116/hr 1-year Reserved. Free Tier for up to 12 months.

Amazon RDS is AWS's managed relational database service for MySQL, PostgreSQL, MariaDB, Oracle, SQL Server, Db2 and Aurora, billed by the instance-hour.

17 /18 features Full profile

Cloud SQL

by Google

No reviews yet

Pricing: Usage billed per hour: shared-core from $0.0105, Enterprise $0.0413 per vCPU plus memory, Enterprise Plus $0.054 per vCPU; storage, egress and SQL Server licences extra. $300 trial credit.

Cloud SQL is Google Cloud's managed relational database service for MySQL, PostgreSQL and SQL Server, billed by usage for teams building on Google Cloud.

10 /18 features Full profile

Feature Comparison

Relational Databases features side-by-side

Feature	Amazon RDS 17/18 features	Cloud SQL 10/18 features
ACID transactions with row-level locking Multi-statement transactions commit or roll back atomically while concurrent writers lock individual rows, not whole tables.	Yes	No
Declarative table partitioning Large tables split by range, list or hash key, with the planner pruning partitions that a query cannot touch.	Yes	No
Native JSON column type with indexing A binary JSON type that can be queried by path and indexed, rather than JSON stored as plain text.	Yes	No
Built-in full-text search indexes Text columns indexed for ranked keyword and phrase search without deploying a separate search engine.	Yes	Yes
Streaming physical replication to standbys Hot standby replicas receive the write-ahead log continuously and can serve read-only queries.	Yes	Yes
Logical replication of selected tables Row changes published per table or schema to subscribers, including subscribers on a different major version.	Yes	Yes
Synchronous multi-replica commit A write is acknowledged only after one or more standbys confirm it, so a failover loses no committed data.	Yes	Yes
Automatic failover to a standby The engine or its bundled cluster manager detects a failed primary and promotes a replica without an operator.	Yes	Yes
Point-in-time recovery from archived logs Restore the database to any moment between backups by replaying archived transaction logs.	Yes	Yes
Online schema changes without table locks Columns can be added and indexes rebuilt while the table stays readable and writable.	No	No
Columnstore indexes for analytics Column-oriented storage on selected tables or indexes to speed aggregate queries over large row counts.	Yes	No
Row-level security policies Per-row read and write predicates that the engine enforces for each database role.	Yes	No
Transparent data encryption at rest Data files and logs encrypted on disk by the engine itself, with keys held outside the data directory.	Yes	Yes
Stored procedures in multiple languages Server-side procedures written in SQL plus at least one general-purpose language such as Python or Java.	Yes	Yes
Extension or plugin architecture Third-party extensions load new data types, index methods or functions into the engine without a fork.	Yes	Yes
Query store or plan history Execution plans and their statistics retained over time so regressions can be found and a plan pinned.	Yes	No
In-place major version upgrade tooling A supported tool upgrades the data directory across major versions without a full dump and reload.	Yes	Yes
Open source core under a permissive license The engine's source is public and the community edition can run in production with no license fee.	Yes	No
Overall Rating	No reviews yet	No reviews yet

Amazon RDS 17/18 features

ACID transactions with row-level locking: Yes
Declarative table partitioning: Yes
Native JSON column type with indexing: Yes
Built-in full-text search indexes: Yes
Streaming physical replication to standbys: Yes
Logical replication of selected tables: Yes
Synchronous multi-replica commit: Yes
Automatic failover to a standby: Yes
Point-in-time recovery from archived logs: Yes
Online schema changes without table locks: No
Columnstore indexes for analytics: Yes
Row-level security policies: Yes
Transparent data encryption at rest: Yes
Stored procedures in multiple languages: Yes
Extension or plugin architecture: Yes
Query store or plan history: Yes
In-place major version upgrade tooling: Yes
Open source core under a permissive license: Yes

No reviews yet

Cloud SQL 10/18 features

ACID transactions with row-level locking: No
Declarative table partitioning: No
Native JSON column type with indexing: No
Built-in full-text search indexes: Yes
Streaming physical replication to standbys: Yes
Logical replication of selected tables: Yes
Synchronous multi-replica commit: Yes
Automatic failover to a standby: Yes
Point-in-time recovery from archived logs: Yes
Online schema changes without table locks: No
Columnstore indexes for analytics: No
Row-level security policies: No
Transparent data encryption at rest: Yes
Stored procedures in multiple languages: Yes
Extension or plugin architecture: Yes
Query store or plan history: No
In-place major version upgrade tooling: Yes
Open source core under a permissive license: No

No reviews yet

Pricing

Amazon RDS

Usage-based, per instance-hour. MySQL or PostgreSQL on db.t4g.micro in us-east-1: $0.016/hr On-Demand, $0.0116/hr 1-year Reserved. Free Tier for up to 12 months.

See all plans

Cloud SQL

Usage billed per hour: shared-core from $0.0105, Enterprise $0.0413 per vCPU plus memory, Enterprise Plus $0.054 per vCPU; storage, egress and SQL Server licences extra. $300 trial credit.

See all plans

Explore Each Product

Amazon RDS Read reviews, features & pricing

Cloud SQL Read reviews, features & pricing

Looking for more options?

Browse all Relational Databases