With an average of 15M transactions processed annually, our client is a prominent payment processor in the USA. All payment processing operations, including underwriting, onboarding, risk assessment, billing, and merchant funding, are conducted in-house.
The cornerstone of their success is the longstanding relationships they have built through decades of work in payment processing.
The Challenge
Meeting the client’s demands proved to be a multifaceted challenge. Their stringent requirements included a fixed processing duration for payment transactions, mandating accuracy down to the minutest detail. This presented a complex hurdle, as even the smallest deviation from perfection could lead to financial discrepancies and user dissatisfaction.
In addition to these precision demands, the client also expressed their need for a scalable architecture that could handle a tenfold increase in transaction volume. Furthermore, with the ongoing concerns over cloud expenditure, the client expected a significant reduction in cloud costs while delivering uncompromising service quality. Balancing these intricate demands while maintaining peak performance was the challenge that lay before us.
Navigating this intricate web of client expectations, accuracy requirements, scalability, and cost efficiency presented a formidable test. It demanded a comprehensive approach and innovative solutions to not only meet but exceed these expectations.
What did
Intronsoft do
Prior to this project, client had already been collaborating with Intronsoft for more than four years. In order to manage the scaling with the new architecture, a fresh team was assembled.
To accomplish this task, Intronsoft designated a team consisting of a System Architect, two Back-end Developers, two Front-end Developers, a DevOps Engineer, a QA Engineer, and a Project Manager.
Leveraging AWS’s versatile cloud infrastructure, we deployed all applications as containers in ECS Fargate, a highly scalable container service. This architecture allowed us to meet the client’s requirement for both horizontal and vertical scalability, ensuring seamless adaptability to changes in transaction volume.
To enhance system efficiency, we preloaded essential data, including lookup information, settings, and aggregated historical data, into Redis Cache. This not only accelerated data retrieval but also reduced the load on other components of the system. We incorporated a strategic data processing approach, applying business logic to batch data that flowed between Java and the Redis Cache layer. This enabled real-time, accurate data processing and contributed to the remarkable precision that the client demanded.
The final piece of our solution was the integration of AWS Opensearch Database, where the fully processed data was meticulously stored. Notably, we harnessed the versatility of Redis Cache as a Message Queue, effectively eliminating the need for a separate queue infrastructure. This streamlining not only reduced system complexity but also led to cost savings, aligning with the client’s objective of minimizing cloud expenditure.
The Results
- 2 million transactions in just 16 minutes, enhancing efficiency.
- Monthly cloud costs stay under $2,000 with 10X scalable infrastructure.
- Achieved and surpassed penny-level accuracy in transactions.
- Delivered these results within 3 months of project initiation.
- Automated deployments and scaling for future sustainability and growth.
