OPPORTUNITY
A large financial services company purchased a 3rd party ESOP projection platform they planned to offer as a SAAS offering. This solution contained a 20-year-old calculation engine written in a language with little to no support at the firm. They needed the ability to scale this platform based on usage and on board new customers to the platform with ease.
SOLUTION
The decision was made to use AWS Aurora Serverless as the current solution was using a Postgres database. This would allow the organization to scale up and down based on usage without the need to manage infrastructure.
The front-end of the application was written in PHP and the calculation engine was written in C++. Both of these were containerized so that dependencies could be managed with ease.
The PHP and C++ code both expected to have a disk volume exposed to the database engine so that copy commands could be executed to load data. Additionally, we removed the copy calls and used the AWS PHP and C++ SDKs to share data via an S3 bucket, and also changed the copy commands to be table_import_from_s3 statements.
RESULTS AND IMPACT
By implementing this solution the company was able to reduce compute costs on the mainframe and provide some headway for future growth without additional costs. By taking advantage of ECS Fargate the financial services organization was able to scale up and down their solution based on demand, providing significant cost savings.