success story

A leading and multinational government administration services company saved an estimated 45% in operations costs by migrating to AWS

The client operated a separate instance of the enrollment broker application for each territory across North America, Australia, the UK, and KSA.  Thus, creating siloed processes run by heterogeneous data. The client, therefore, sought to transform its applications to enhance batch performance as it was adding to the cost of its year-on-year operation

The client partnered with Virtusa, banking on our experience and expertise to validate the technical approach, pilot the first migration, and save an estimated 45% in operations costs.

The Challenge

The current enrollment offering systems/processes lacked scalability. Any increase in daily enrollments significantly slowed batch processing and needed extra time to monitor and ensure no batch failures. The client’s existing application instance comprised 60% common codebase and 40% state-specific customizations. Thus, the Oracle database was customized with heavy code objects and user-defined functions. Moreover, its PostgreSQL did not have packages, and hence converting the Oracle packages required significant efforts. All of this was leading to a rise in the client’s already high year-on-year operational cost.

The Solution

The client partnered with Virtusa to leverage the benefits of comprehensive AWS autoscaling instances. We deployed three agile teams to work parallelly for schema conversion, data migration, post-migration validations, and reconciliations.

Oracle to Amazon Aurora PostgreSQL migration:

The heterogeneous database migration encompassed two aspects: To convert the source scheme to target DB and to migrate the data to target DB.

We automated remediation of PostgreSQL compatibility issues using a regex-based remediation tool and built a framework for schema recon and post-migration data validations. CI/CD  pipeline deployment accelerated migration efforts while monitoring Oracle performance metrics against Aurora PostgreSQL metrics.

Later, we extracted user data from the source database to a CSV file to be loaded into a temp table in PostgreSQL. Our team also created a SQL statement to retrieve data from the temp table in CREATE USER statements and ran it to set up users in PostgreSQL.

Application migration

Once the client’s existing Oracle Database was re-architected to Amazon Aurora, its application servers, web servers, and ETL servers were re-hosted to EC2s. We remediated all Oracle-specific syntaxes and data types in the application codebase. We also created blueprints for infra provisioning:

  • iBatis framework leveraged as an ORM layer between database and Java objects
  • Kettle used for enrollment batch processing
  • Automated remediation of Oracle-specific code in iBATIS and Kettle jobs using a regex-based remediation tool handled
  • Changes to SQL statements, jdbcTypes, and type Handlers in iBatis Mapper files
  • Datatype changes in Java Objects
  • SQL statements in Kettle jobs
The Benefit

Virtusa helped the client implement a scalable factory model for migrating the rest of the enrollment broker applications to:

  • Save an estimated 45% in operations costs
  • Reduce manual efforts by 70%
  • Improve batch processing performance
  • Streamline run time for any given load volume
  • Reuse for all future migrations as well
Cloud migration services

Rapidly and successfully migrate your applications, data, and infra to the cloud with minimum business disruption.

Related content