Driving large-scale cloud migration success with zero downtime
About the company
A prominent U.S.–based financial institution.
Challenge
The financial institution had multiple critical applications on more than 200 relational databases. It was looking to migrate these relational databases from Oracle to Amazon Web Services Relational Database Service (AWS RDS) for PostgreSQL and subsequently improve performance, governance and scalability. These are critical no-downtime applications integral to market functioning, compliance and integrity.
While migrating to AWS RDS, it was also important to establish data governance for a Relational Database Management System (RDBMS) and meet the Payment Card Industry, Sarbanes-Oxley Act and Nonpublic Personal Information/Personally Identifiable Information compliance requirements. Another critical goal was to enable the high availability of apps and scalability of relational databases based on the needs and traffic requirements of the applications.
What we did
- Data architecture
- Data governance
- Data encryption using AWS Key Management Service (KMS) setup
- Multi-Availability Zone (Multi-AZ) disaster recovery setup
- Data catalog
- Schema migration
- AWS RDS migration
Solution
The company had already adopted the AWS cloud and had mature enterprise operations, governance, DevOps and approval policies. Our team utilized enterprise policies, foundations and frameworks to establish the following:
- Provided data governance
- Set up data encryption using KMS
- Set up Multi-AZ disaster recovery
- Reengineered and developed pipelines and migrated data from on-premises Oracle relational databases to RDS for PostgreSQL
- Set up the application testing suite for execution and a Jenkins Pipeline to run and manage automated testing
- Created a schema in the target PostgreSQL database by converting the source Oracle database schema utilizing the AWS Schema Conversion Tool
- Set up AWS Database Migration Service (DMS) to load data from the source to the target database and implemented automated processes to trigger DMS
- Utilized open-source packages like Ora2Pg to convert Oracle and MySQL databases to PostgreSQL automatically; enabled the pub/sub model with a relational data warehouse to sync the application’s source data
Result and outcomes
The large-scale migration of an RDBMS to a cloud-native Amazon RDS helped the company achieve the following outcomes and benefits:
Zero
downtime during the large-scale migration
45+
application relational databases migrated to the cloud
5,000+
total active users migrated seamlessly without disrupting the end-user experience
Business benefits
- Considerable savings in database licensing costs
- Better scalability, resiliency and governance
- Faster time to market with increased speed of changes
Technical benefits
- Change data capture enabled for the application data warehouse to application database sync
- Zero disruption to critical applications during service level agreement hours