Ascend Money Improves Application Performance by 40% Using AWS Graviton–Based EC2 Instances

Based in Thailand, Ascend Money provides financial services to 50 million people throughout Southeast Asia, with a mission of facilitating access to innovative financial services, leading to better lives. Its flagship product, TrueMoney Wallet, is an app that provides digital financial solutions, such as cashless shopping, bill payment, lending, high yield saving, investment, insurance purchasing, and other services. The company began a partial migration to the cloud in 2014 while keeping some of its workload on premises due to local regulations. It orchestrates most of its cloud-based workload using Amazon Elastic Kubernetes Service (Amazon EKS), a managed container service to run and scale Kubernetes applications. In conjunction with Kubernetes Cluster Autoscaler, Ascend Money uses AWS Auto Scaling, which monitors a company’s applications and automatically adjusts capacity to maintain steady, predictable performance at low cost. As a result, Ascend Money can provision and automatically scale Amazon EKS nodes across many different instance types. To reduce overall compute costs, Amazon EC2 Spot Instances provide unused Amazon EC2 capacity in the cloud for up to a 90 percent discount compared to Amazon EC2 On-Demand Instances, for which companies pay for compute capacity by the second with no long-term commitments. Spot Instances work well for time-flexible workloads, such as Ascend Money’s testing and development applications. By adopting Spot Instances and AWS Savings Plans, a flexible pricing model offering lower prices compared to On-Demand pricing in exchange for a specific usage commitment, Ascend Money saved 70 percent on compute costs.

In 2019, AWS introduced the AWS Graviton2 processor, designed by AWS to deliver optimal price performance for cloud workloads running in Amazon EC2 and which uses ARM64 architecture instead of x86. Using AWS Graviton2 processors and the versatile x86 processors it was using for Spot Instances, Ascend Money built an innovative multi-arch approach that involves running services designed for one computer architecture on another. Ascend Money developed its compute strategy alongside AWS compute specialists, an AWS account team, and AWS Enterprise Support, which provides concierge-like service where the main focus is helping companies achieve their outcomes and find success in the cloud. The strategy involved expanding the use of Spot Instances using x86 processors for fault-tolerant and scalable workloads while using On-Demand Instances on AWS Graviton2 processors. This combination offered desirable cost and price performance across the landscape of compute on AWS.

Because companies in the financial services industry face strict regulations regarding anti-money-laundering activities, Ascend Money must keep a close eye on its 7 million daily transactions. For name screening and transaction monitoring, Ascend Money uses Amazon EMR, a cloud big data solution for running large-scale distributed data processing jobs, interactive SQL queries, and machine learning applications using open-source analytics frameworks. “It’s very simple for us to manage and operate,” says Peerawit Phuangkaeo, head of technical operations at Ascend Money. Beginning with Amazon EMR, Ascend Money began migrating some of the production workload to AWS Graviton2 processors in 2021. “On average, we see 30–40 percent better performance for our workload and total cost savings around 20 percent per comparable x86 processor–based instance, without any code change,” says Phuangkaeo. “I have been encouraging our developers to use AWS Graviton2 processors. They are confident when they try them and see the results.”

Ascend Money also runs its mobile gateway application using AWS Graviton2 processors. “Our team didn’t have to fine-tune anything applying AWS Graviton2 processors, and our performance was about 20 percent better,” says Phuangkaeo. “Using AWS Graviton2 processors helps us think about new services and innovation rather than performance issues.” For example, developers have been working on improving optical character recognition, which recognizes the text in scanned documents and, therefore, eases customer registration.

Ascend Money continues to use Spot Instances to control the costs of its testing environment. The company addressed performance challenges with the TrueMoney Wallet when it was testing potential bottlenecks, resulting from an unpredictable influx of users. By using Spot Instances, Ascend Money saved about 60 percent in costs while helping developers become confident enough in the app’s performance to bring new features online.

The company built its continuous integration and continuous delivery pipeline to support running the two types of instances—those running using x86 processors and those using AWS Graviton2 processors—in parallel. Developers can choose the instance before building the code for a particular use case, and they have further flexibility because AWS Graviton2 processors are compatible with many programming languages, including Golang, NodeJS, Java, and Python. “We are reducing the effort for the developer,” Phuangkaeo says. “We can reduce the time to launch new services in production. Using AWS, we know that we have quality and security in the pipeline.”

Ascend Money continues to launch new services for customers, such as a lending business to purchase motorcycles in Vietnam. To support its additional digital services, Ascend Money is expanding its use of both types of instances and is working alongside AWS to find ways to adopt Spot Instances in its production workload.

Because Ascend Money has kept about 30 percent of its workload on premises, it has begun to explore the use of Amazon EKS Anywhere, a new deployment option for Amazon EKS that helps customers create and operate Kubernetes clusters on customer-managed infrastructure, supported by AWS. “Using AWS, we have gained better performance,” says Phuangkaeo. “The direct effect to our business is that we can deliver more new innovative services to our customers.”