AWS Database Migration Service Specialized2016年〜
A fully managed migration service that moves databases to AWS with minimal downtime
What It Does
AWS Database Migration Service (DMS) is a fully managed service for migrating databases from on-premises or other clouds to AWS. It supports both homogeneous migrations (Oracle to Oracle) and heterogeneous migrations (Oracle to Aurora). The source database continues operating during migration, minimizing downtime. Its continuous data replication (CDC) feature keeps source and target in sync even after the initial migration, enabling phased migration strategies.
Use Cases
It is used for migrating on-premises Oracle or SQL Server to Aurora or RDS, reducing license costs by moving from commercial to open-source databases, migrating data warehouses to Redshift, creating database replicas for dev/test environments, consolidating databases in multi-cloud environments, and feeding real-time data replication to analytics platforms - covering a wide range of database migration and synchronization scenarios.
Everyday Analogy
Think of it like moving fish between aquariums. When transferring fish (data) from an old tank (source DB) to a new tank (target DB), moving them all at once is stressful. DMS gradually transfers the water to the new tank (continuous replication) while keeping the fish healthy (no downtime). It even works when the tanks are different shapes (heterogeneous DBs).
What Is DMS?
AWS Database Migration Service is a fully managed service that greatly simplifies database migrations. Traditional database migrations required manually managing many steps - schema conversion, data transfer, integrity verification, and application cutover. DMS automates data transfer and continuous replication, reducing the risk and effort of migration projects. It supports over 20 database engines including Oracle, SQL Server, MySQL, PostgreSQL, MongoDB, and Amazon Aurora.
How Migration Works
The DMS migration process consists of three steps: creating a replication instance, configuring source and target endpoints, and defining a migration task. By combining full load (bulk transfer of all data) with CDC (change data capture), changes to the source database continue to be replicated to the target in real time after the initial data migration. Table mapping rules let you target specific schemas or tables for migration, and apply column name transformations or filtering.
Integration with Schema Conversion Tool
For heterogeneous database migrations, integration with AWS Schema Conversion Tool (SCT) is essential. SCT automatically converts the source database schema (table definitions, stored procedures, views, etc.) to the target database format. Items that cannot be automatically converted are output as a report, clearly identifying what needs manual attention. The two-step approach of converting the schema with SCT and then migrating the data with DMS enables systematic heterogeneous database migrations. For a comprehensive understanding of Schema Conversion Tool integration, specialized books (Amazon) are also a helpful reference.
Migration Validation and Monitoring
DMS includes data validation features that automatically verify that source and target data match. It compares row counts and validates data content, recording any discrepancies in a report. Migration task progress can be monitored in real time via CloudWatch metrics, including replication lag, error counts, and throughput. You can set up SNS notifications to receive immediate alerts when issues occur.
Things to Watch Out For
- Heterogeneous database migrations require schema conversion with Schema Conversion Tool, and manual conversion of stored procedures may be needed
- Size the replication instance appropriately based on data volume and change frequency to prevent performance degradation during migration
