Amazon Database Migration Service

Amazon Database Migration Service (AWS DMS) moves relational databases, data warehouses, NoSQL databases and other data stores to AWS or between cloud and on-premises setups. It supports one-time migrations as well as ongoing replication to keep source and target data in sync, including moves between like database engines or different ones. AWS says more than 20 databases and data warehouses are supported as endpoints. DMS Schema Conversion assesses and converts schemas and most database code objects, such as views, stored procedures and functions, and identifies objects that need manual conversion. DMS Serverless provisions and scales migration capacity according to transaction volume, with capacity billed hourly. The Multi-AZ option uses redundant replication instances, and a migration can resume after an interruption. DMS integrates with CloudTrail and CloudWatch Logs; its API and AWS CLI can automate replication task management and custom monitoring. Security options include TLS, access policies and VPC endpoints. At least one replication endpoint must be an Amazon RDS or EC2 database; on-premises-to-on-premises replication is not supported. The free tier offers $100 in credits, with up to an additional $100 for activating foundational AWS services; credits are valid for up to 12 months.

AWS DMS covers varied migration paths and ongoing replication, with schema conversion and recovery options. Its endpoint requirement is important: it cannot replicate solely between on-premises databases.

AWS DMS suits teams migrating databases or analytics workloads to AWS, including moves from self-managed databases to AWS-managed services. It can also support ongoing replication where at least one endpoint is Amazon RDS or an EC2-hosted database.

Amazon Database Migration Service (AWS DMS) is a managed service for moving relational, warehouse, NoSQL, and other data stores into AWS or between cloud and on-premises environments. It is best for teams modernizing databases for AWS while needing either a one-time cutover or a synchronized transition. The service is capable and resilient, but its endpoint rule makes it unsuitable for purely on-premises replication.

Overview

AWS DMS supports homogeneous migrations between matching database engines and heterogeneous migrations between different engines. It can perform a single migration or continuously copy changes so source and target remain aligned. AWS says more than 20 databases and data warehouses can serve as source or target endpoints, a useful breadth for organizations consolidating mixed estates.

This is a SaaS deployment exposed through web and API interfaces. Discovery and assessment, database migration, and continuous replication are all supported. The important boundary is architectural rather than cosmetic: for replication, the source or target, or both, must be Amazon RDS or an EC2-hosted database. On-premises-to-on-premises replication is not supported, so DMS is a poor fit when AWS is not part of the topology.

Key features

Migration and schema work

DMS Schema Conversion evaluates schemas and converts most database code objects, including views, stored procedures, and functions. It also identifies objects requiring manual conversion. That combination can shorten planning and expose incompatibilities early, while still leaving engineering work for database-specific code that cannot be converted automatically.

Elastic capacity and recovery

DMS Serverless automatically provisions and scales migration capacity with transaction volume. It removes the need to select replication capacity in advance, but hourly charges vary with data volume and the maximum DCU setting, so a high ceiling can affect operating cost. On-demand replication instances provide a more explicit instance choice and can resume a migration after an interruption. Multi-AZ deployment adds redundant replication instances for high availability, which is valuable for production cutovers and long-running replication rather than short, disposable jobs.

Operations and security

CloudTrail and CloudWatch Logs integrations support audit trails and operational monitoring. The API and AWS CLI can automate replication-task management and custom monitoring, making DMS a stronger match for teams already operating AWS through scripts and pipelines. API clients must use TLS 1.2, with TLS 1.3 recommended; resource-based access policies and AWS PrivateLink VPC endpoints add control over access paths. Endpoint connections can use SSL/TLS, replication instances can run inside a VPC, and specified sensitive columns can be masked.

Pricing

AWS Free Tier

The AWS Free Tier costs nothing and offers $100 in credits, with up to an additional $100 for activating foundational AWS services. Credits are valid for up to 12 months. When credits are depleted, a Paid Plan is required to continue using DMS. Free Tier credits are unavailable in AWS GovCloud (US) and China.

AWS DMS Serverless

Custom pricing applies, billed per hour according to capacity used in DCUs. The rate changes with data volume and the maximum DCU setting. It suits intermittent or variable workloads that benefit from automatic scaling, but budgeting requires attention to both workload size and the configured ceiling.

On-demand instances

Custom pricing is billed hourly with no long-term commitment. Cost varies by instance type and AWS Region, and additional log storage may cost extra. This is the flexible choice for migrations whose duration or capacity is uncertain, provided regional instance and storage charges are accounted for.

Database Savings Plans

Custom pricing requires a one-year commitment to a specified usage amount measured in dollars per hour for eligible AWS DMS usage. Consistent usage can save up to 20%. The plan is aimed at predictable, sustained replication; it gives up the flexibility of purely on-demand billing and risks paying for committed usage if migration activity falls.

Platforms

AWS DMS is available through web and API interfaces. Its API and AWS CLI support automated task administration and monitoring, while replication instances and endpoint connections can be placed within AWS networking controls such as a VPC.

Who it's for

DMS is a strong choice for organizations moving self-managed databases to managed AWS databases, modernizing analytics workloads, or keeping a live copy during a staged cutover. It also fits AWS operations teams that need audit logging, scripted control, private connectivity, and Multi-AZ resilience. Choose another tool when both endpoints must remain on premises, when a migration does not involve RDS or EC2, or when a one-off schema conversion is the only requirement.

Pros and cons

• Pros: One-time migration and continuous replication cover both cutovers and synchronized transitions.
• Pros: Support for more than 20 database and warehouse endpoints helps mixed-engine estates.
• Pros: Schema Conversion handles many code objects and flags manual work instead of hiding gaps.
• Pros: Serverless scaling, Multi-AZ redundancy, and interruption recovery address changing load and operational risk.
• Pros: TLS, VPC placement, column masking, PrivateLink, CloudTrail, and CloudWatch support controlled deployments.
• Cons: Replication cannot run solely between on-premises databases because at least one endpoint must be on RDS or EC2.
• Cons: Serverless and on-demand costs vary with capacity, volume, region, instance type, and possible log storage.
• Cons: Schema Conversion still requires manual handling for flagged objects, so it does not eliminate database remediation.

Alternatives

For a broader shortlist of infrastructure migration products, see Cloud Migration Tools. For products centered on keeping datasets synchronized, see Data Replication Software.

• Azure Application Insights is a paid, web-and-API-capable option with a free plan and usage-based Azure Monitor Logs pricing; choose it when Azure monitoring is the priority rather than AWS database migration.
• pgloader is a freemium, Linux self-hosted tool with a free Community plan; choose it for a focused PostgreSQL-oriented, self-managed migration.
• Striim is a freemium platform with a free trial, API, web, and self-hosted options; choose it when event-pipeline packaging and its Developer or Cloud offerings fit better.
• AWS Schema Conversion Tool is free for Windows and Linux; choose it when schema assessment and conversion are needed without DMS replication.
• AWS Transform MGN (formerly AWS Application Migration Service) offers a free period of 2,160 hours per source server; choose it for server workload migration rather than database replication.
• Google Cloud Migration Center is free for cost estimates and assessments; choose it for Google Cloud planning before committing to migration services.
• Google Migrate to Containers is free for migrating workloads to Google Cloud; choose it for containerizing workloads rather than moving database data.
• Ora2Pg is free GPL v3 software; choose it for an Oracle-to-PostgreSQL-focused, self-managed conversion.

Verdict

Choose AWS DMS when your migration includes AWS RDS or EC2 and you need broad engine coverage, ongoing replication, and AWS-native resilience and controls. Look elsewhere if you require on-premises-to-on-premises replication or want a narrowly focused, free schema-conversion utility without a migration service.