Trust Bank Builds a Scalable and Innovative Digital Bank on AWS

When Trust Bank launched in 2022, Singapore was arguably oversaturated with banking options, with 98 percent of the population having at least one bank account. Yet, Trust Bank reached more than 100,000 customers within 10 days of its launch. In its first year, it onboarded around 14 percent of the market, making it one of the world’s fastest growing digital banks.

New user setup was setting benchmarks, taking less than 3 minutes and with a straight-through processing rate higher than 90 percent. Trust also rolled out a broad range of products from the start, including current deposit accounts, credit and debit cards, and an extensive rewards platform. Few banks achieve such a well-developed customer  proposition this quickly. To support this rapid growth, the bank was built entirely on Amazon Web Services (AWS).

Six months after its launch, Trust Bank decided to improve resiliency and increase input/output (I/O) performance by migrating from Amazon RDS for PostgreSQL—which gives companies access to the familiar PostgreSQL database engine—to Amazon Aurora, which provides unparalleled high-performance and availability at global scale with full MySQL and PostgreSQL compatibility, at 1/10th the cost of commercial databases.

By adopting AWS Graviton processors, which provide up to 40 percent better price performance compared with the corresponding x86-based instances, Trust Bank reduced database costs by 18 percent, even with 30 percent more customers. It also reduced the I/O cost on Amazon Aurora and increased the maximum I/O operations per second by 10 times. Now, P99 latency for payment processing is 67 percent lower.

Trust Bank’s data lake stores hundreds of terabytes of data on Amazon Simple Storage Service (Amazon S3), an object storage built to store and retrieve any amount of data from anywhere. Data lake inputs stream in near real time from Apache Kafka—an open-source distributed event streaming system—to Amazon S3 using Apache Kafka Connect. Data is then used for business intelligence and machine learning (ML) use cases.

Trust Bank implemented a data mesh model to manage data governance and data permissions. In addition, all sensitive data are tokenized in the data lake.

Trust Bank is deploying ML to enhance operations and provide personalized customer recommendations using Amazon SageMaker, a fully managed ML service. It’s also using Amazon EMR, a cloud big data solution, for training and inference. “We’re innovating and releasing new products much faster due to our life cycle on AWS,” says Rajay Rai, chief information officer at Trust Bank.