Background

Oracle Database is one of the most widely used relational databases in enterprise applications, but its high licensing fees and operational costs place a heavy burden on organizations. As cloud computing becomes mainstream, more and more enterprises are choosing to migrate their Oracle databases to the AWS cloud platform to reduce costs and improve elasticity and availability.

This article describes a practical case where UltraPower Cloud Business assisted an enterprise client in migrating their Oracle database to AWS, covering key aspects including migration solution design, tool selection, migration implementation, and validation.

Customer Pain Points

• High licensing costs: Oracle Database licensing fees are expensive, consuming a significant portion of the annual IT budget.
• Complex operations: Self-managed databases require a dedicated DBA team, resulting in high labor costs.
• Poor scalability: Traditional physical server expansion cycles are long and difficult to handle business peak loads.
• Weak disaster recovery: Existing DR solutions are complex, and RTO/RPO metrics are difficult to meet business requirements.

Migration Solution

Target Architecture

Based on the client's business characteristics and data volume, the following migration paths were defined:

• Homogeneous migration: Migrate Oracle to Amazon RDS for Oracle, retaining original Oracle features and reducing application transformation costs.
• Heterogeneous migration: Migrate non-core business Oracle databases to Amazon Aurora PostgreSQL, eliminating Oracle dependency and significantly reducing licensing fees.

Migration Tools

The following AWS native tools were primarily used for this migration:

• AWS Schema Conversion Tool (SCT): Used to assess the migration complexity of database objects and automatically convert DDL, stored procedures, functions, and other database objects.
• AWS Database Migration Service (DMS): Used for full data migration and continuous incremental synchronization (CDC) to achieve minimal downtime migration.
• AWS DMS Fleet Advisor: Used to automatically discover and assess the source database environment and generate migration recommendation reports.

Migration Implementation Steps

Phase 1: Assessment and Planning

1. Use AWS SCT to scan the source Oracle database, generate an assessment report, and identify incompatible objects.
2. Analyze the migration complexity of data volume, table structures, stored procedures, triggers, and custom functions.
3. Develop a detailed migration plan including migration sequence, time windows, and rollback procedures.
4. Set up the target database environment on AWS (RDS/Aurora) and complete network, security group, and parameter group configuration.

Phase 2: Schema Conversion

1. Use AWS SCT to automatically convert compatible database objects (tables, indexes, views, etc.).
2. For stored procedures, triggers, and other objects that SCT cannot automatically convert, the DBA team performs manual transformation.
3. Execute the converted DDL scripts on the target database to complete schema creation.

Phase 3: Data Migration

1. Configure the AWS DMS replication instance and set up source (Oracle) and target (RDS/Aurora) connections.
2. Start the full migration task to migrate historical data to the target database.
3. After full migration is complete, enable CDC (Change Data Capture) mode to continuously synchronize incremental data.
4. Monitor DMS task status to ensure data latency is within acceptable limits.

Phase 4: Validation and Cutover

1. Perform row count and checksum comparisons on migrated data to ensure data integrity.
2. Run the application in a test environment to verify that business functions work correctly.
3. Execute the cutover during a low-traffic period, redirecting application connections to the target database.
4. Continuously monitor database performance and application status after cutover to confirm successful migration.

Migration Results

• Cost reduction: By migrating to Aurora PostgreSQL, Oracle licensing fees were reduced by approximately 70%, and overall database operating costs decreased by more than 40%.
• Performance improvement: Amazon Aurora's read/write performance showed significant improvement compared to the original Oracle environment, with query response times reduced by approximately 30%.
• High availability: Amazon RDS/Aurora provides Multi-AZ deployment with automatic failover, reducing RTO from hours to minutes.
• Simplified operations: Managed database services automatically handle backups, patching, monitoring, and other operational tasks, freeing the DBA team from routine maintenance to focus on business optimization.
• Minimal downtime: With DMS CDC functionality, business cutover downtime was controlled to within 30 minutes.

Key Takeaways

Oracle database migration is a complex systems engineering project that requires thorough upfront assessment and planning. Key lessons from this migration include:

• Identify incompatible objects early and allow sufficient time for transformation to avoid handling issues during the migration window.
• Leveraging AWS SCT and DMS tools can significantly reduce migration workload and risk.
• Develop a comprehensive rollback plan to ensure rapid recovery if issues arise during migration.
• Conduct thorough data validation before and after migration to ensure data integrity and consistency.
• Choose an appropriate migration time window and execute cutover operations during business off-peak hours whenever possible.

For more information about Oracle database migration to AWS, please contact the UltraPower Cloud Business team. We provide professional migration assessment and implementation services.

Back to Tech Blog