Faez Hasnan

Interest Calculation for an Account Management System

Introduction

This case study highlights the successful implementation of an interest calculation system within an internal platform managing 1 million accounts, spanning 100 million data points. The system accurately performs daily, monthly, quarterly, and yearly interest calculations while accommodating additional functionalities such as bonus interest, backdated transactions, and day count calculations. Leveraging .NET and the Task Parallel Library (TPL), we developed a highly efficient and scalable service, containerized using Docker. The service also served as the foundation for various internal special projects. Processing wise, the interest calculation process was significantly improved, reducing the time required from hours to minutes.

Problem Statement

The platform faced the challenge of managing a large expanse of accounts while ensuring precise and timely interest calculations. With an extensive dataset of 100 million balance mutations, the existing manual calculation methods were time-consuming. We sought an automated solution that would not only optimise interest calculations but also incorporate advanced features and handle the scaling demands.

Solution

To address the challenges, we employed a comprehensive software solution using .NET and the Task Parallel Library (TPL). These technologies provided the necessary performance, scalability, and flexibility to handle complex interest calculations and integrate additional functionalities.

Architecture and Implementation

The interest calculation system comprised the following essential components.

Data Management

A centralized database was implemented to manage account details, financial information, and historical transaction data. This database served as the primary data source for the interest calculation system.

Interest Calculation Engine

The core of the application was the interest calculation engine, built using .NET and harnessing the parallel processing capabilities of TPL. The engine encompassed advanced algorithms capable of performing periodic interest calculations accurately. It also incorporated provisions for bonus interest calculations, handling backdated transactions, and day count calculations to meet the extended requirements of clients.

Task Parallelism

By leveraging TPL, the interest calculation tasks were partitioned into smaller units and distributed across multiple threads and processors, enabling parallel processing and efficient utilization of system resources. This approach significantly reduced the overall execution time, leading to faster and parallel calculations of smaller data points.

Dockerization

The entire application was containerized using Docker, ensuring seamless deployment and consistent operation across various environments. Dockerization simplified the deployment process, enhanced portability, and eliminated compatibility issues.

Results and Benefits

The implementation of the interest calculation system using .NET, TPL, and Docker resulted in significant benefits.

Drastic Reduction in Processing Time

The new interest calculation system revolutionized the efficiency of interest calculations, reducing the processing time from the magnitude of hours down to mere minutes. The parallel processing capabilities of TPL enabled efficient distribution of tasks, resulting in faster completion and improved operational efficiency.

Accuracy and Flexibility

Automating the interest calculation process eliminated manual errors, ensuring precise interest calculations of 1 million accounts in total.

Scalability and Performance

The application demonstrated excellent scalability, effortlessly managing 10 million accounts when tested for load and scale. The parallel processing capabilities of TPL enabled the system to handle increased workloads while maintaining optimal performance, enabling future client onboarding without compromising on processing time.