AsiaYo Expands Room Booking Platform on AWS

AsiaYo expanded its Taiwan headquarters to six countries without sending any staff abroad and has saved 50 percent on the time it takes to make architecture changes by using AWS. AsiaYo is a leading platform for booking exclusive accommodations in 60 cities across Asia at competitive prices. The startup uses Amazon Aurora to store customer reservation and payment data, Amazon ElastiCache for Redis to handle spikes in application traffic, and Amazon CloudWatch for application and infrastructure monitoring. It also uses AWS Management Console to easily manage all its AWS services.

Intraregional tourism is on the rise in Asia. According to the World Tourism Association, in 2016 about 80 percent of international arrivals in Asia and the Pacific originated in the same region.

AsiaYo is poised to capitalize on this trend. Founded in 2013 and based in Taipei, AsiaYo lists more than 60,000 rooms and properties across 60 cities on its room booking website. Japan is its largest market outside of Taiwan, and the company also operates in Hong Kong, Malaysia, Singapore, South Korea, and Thailand. Its target audience mostly comprises diverse and intraregional tourists exploring the region. Visitors to the website are guaranteed the most complete customer care experience with support in seven languages and access to exclusive listings at competitive prices.

To ensure a seamless booking experience for its customers at all times, IT infrastructure stability was a top priority. When launching operations, AsiaYo did a cost/benefit analysis comparing Amazon Web Services (AWS) with other cloud providers and found that AWS offered the most benefits at the lowest cost. Specifically, AsiaYo was able to assign just one person to manage its cloud infrastructure on AWS, whereas other providers would have required more people for cloud infrastructure management.

“Using the AWS Management Console, we can easily scale our infrastructure, monitor systems, and balance network loads while lowering our monthly maintenance costs,” says Graffine Wu, RD director at AsiaYo. Wu also notes that compared to other cloud providers, AWS security groups were easy to set up and user-friendly, and they ensured tight control of the company’s network security.

Due to the high volume of business in Japan, AsiaYo chose to run operations out of the AWS Asia Pacific (Tokyo) Region. It has also recently launched business-to-business operations in Greater China, which is able to run from the same Region without any latency concerns. Engineers performed testing for data traveling from its web and mobile applications to servers and saw turnaround times hitting its end-to-end latency target of less than 500 milliseconds.

AsiaYo also launched in Malaysia and Singapore in 2019, running from the AWS Asia Pacific (Singapore) Region to ensure that its data syncs seamlessly across global websites. Nextlink Technology, an AWS Partner Network (APN) Premier Consulting Partner, helped to carry out proof-of-concept testing in Singapore and provided support for billing and accounting, specifically the application of the machine credit plan on the AWS Cloud.

The business uses Amazon Aurora to save customer reservation and payment data, and the team appreciates the speed and simplicity of Amazon Aurora for database management at scale. AsiaYo relies on Amazon DynamoDB for storage and Amazon ElastiCache for Redis for enhancing the extraction and reading performance of databases. It takes advantage of Amazon CloudWatch for application and infrastructure monitoring and Amazon CloudFront for fast content delivery on its website. Developers employ a continuous integration/continuous delivery (CI/CD) model to automate testing and releases, having built the service on Amazon Elastic Compute Cloud (Amazon EC2) instances. The CI/CD monitoring system enables frequent, small-batch releases, which makes it easier for the team to troubleshoot any issues.

In addition to ease of expansion, AsiaYo has benefited from lower maintenance costs on the AWS Cloud, citing savings of more than 50 percent on the staffing required to manage its infrastructure. The business entered six countries without having to send any staff abroad, achieving lower barriers to entry for new travel markets. “We don’t need to build a local data center in every location and hire several employees for maintenance,” Wu says. “Everything we need is right there on the AWS Cloud, and we can modify our system via the AWS Management Console. The simple and user-friendly configuration on AWS Management Console helps make our teams efficient.”

Automation for processes such as CI/CD also facilitate a reduction in labor. Additionally, by using Nextlink’s billing portal, AsiaYo receives a consolidated multinational invoice for its AWS usage, which allows its teams to focus on cost optimization in target areas. It also uses the AWS Total Cost of Ownership (TCO) Calculator to monitor and control monthly spend.

Despite its rapid growth—its compound annual growth rate (CAGR) has hovered around 200 percent over the past five years—AsiaYo has been able to maintain good platform performance since the company’s inception. When assessing stability, AsiaYo is mainly concerned with two aspects: network and storage. “We usually measure stability by application service response time and traffic spike tolerance,” Wu says. “Average response time for our core services is 100–200 milliseconds, which is better than industry averages. Traffic can spike up to 10 times the normal service volume, but we can handle these spikes using redundancy and automatic scaling mechanisms.”

As a startup, AsiaYo has made frequent adjustments to its system architecture to accommodate new markets and features. Management estimates a time savings of 30–50 percent when such changes are made, compared to the time that would have been required with an on-premises data center. “With AWS it’s easy to revamp our architecture,” Wu says. The company maintains release cycles of about three weeks, with a goal of reducing this to two weeks.

One of the factors that differentiates AsiaYo from its competitors is the website’s use of artificial intelligence (AI) to analyze user behavior and develop personalized booking suggestions, what it calls “hyper-personalized journeys.” AsiaYo’s lab is where its AI algorithm AYSort began, and where future AI or machine learning (ML) projects are tested on Amazon EC2 instances. The team is now developing a potential image-recognition tool on its AWS infrastructure, among other ML projects.

In line with its identity as “a tribe of travelers providing accommodations across Asia,” AsiaYo plans to continue developing unique technological capabilities in the travel industry to serve its expanding market.

To learn more, visit aws.amazon.com/products/management-tools.