By migrating to the AWS Cloud, we have reduced our IT costs by at least 20 percent.
Jinwei Lin Assistant Manager, SETTour

Southeast Travel Agency (SETTour), headquartered in Taiwan, is one of the country’s top-five tourism companies. Launched in 1961, the company has branch offices in Taiwan, China, Japan, and the United States. In 2013, SETTour launched its website for customers to review travel destinations and make online travel bookings. Over the last five years, the website has become an increasingly important sales channel for SETTour, generating TWD$30 billion (US$974 million) in annual revenue.

A major challenge was the performance of the on-premises infrastructure behind the SETTour website. The most critical issue was downtime. When the IT team made upgrades to hardware or software, administrators often had to take the website offline. In one instance, website performance was disrupted for two days during a database upgrade. Jinwei Lin, assistant manager at SETTour, says, “It cost us something like TWD$35 million (US$1.13 million) in lost website revenues.”

Another issue with the on-premises infrastructure was around big data. SETTour wanted to analyze large daily volumes of website data to help improve its website’s design. An analysis of traffic—such as how long people stayed on each webpage—would give SETTour valuable information on areas whose design it could improve. Yet, the on-premises infrastructure lacked the compute and storage capacity to provide any kind of big data analysis.

SETTour looked to migrate its on-premises infrastructure to a public cloud service. As Lin explains, “Public cloud infrastructures offer massive amounts of processing power for big data workloads, plus the elasticity to quickly scale up compute and storage capacity.” In his view, a public cloud could also help SETTour improve its website’s uptime. “Cloud operators build a lot of redundancy into their infrastructures to avoid downtime during maintenance work, such as upgrades,” he says.

SETTour assessed offerings from leading cloud-services providers, including the Google Cloud Platform and Amazon Web Services (AWS). “Google Cloud was good, but we felt AWS offered a more comprehensive service. AWS also had great local support through its partner network,” says Lin. With the help of Nextlink Technology, an AWS Partner Network (APN) Premier Consulting Partner, SETTour soon created a migration plan to move its website to the AWS Cloud and developed an environment for big data analysis.

Nextlink also helped SETTour migrate its on-premises web servers to Amazon Elastic Compute Cloud (Amazon EC2), which can be scaled on demand. “Of course, with Amazon EC2, you no longer have to plan for capacity and procure on-premises servers months in advance,” says Lin. Furthermore, Amazon EC2 instances are backed by a service level agreement of 99.999 percent uptime, with rapid failover between instances to maximize uptime. “By switching to the AWS Cloud, our uptimes went from around 95 percent to 99.999 percent. We no longer run the risk of downtime impacting revenues,” says Lin.

SETTour uses Amazon EC2 instances to support its database on Amazon Simple Storage Service (Amazon S3), which SETTour uses to store site images and for data backups. Says Lin, “We find that Amazon S3 meets our needs for flexible and secure storage and helps us get over the issue of storage capacity planning.”

He continues, “By migrating to the AWS Cloud, we have reduced our IT costs by at least 20 percent, because we no longer have to procure and manage an on-premises infrastructure. In turn, this has allowed us to downsize our IT administration team and move team members into development and data analytics roles. We’ve reduced our IT management costs by 50 percent with AWS.”

With the support of Nextlink, SETTour also built its big data environment on the AWS Cloud. Every “click” on a page of its website triggers a serverless-computing AWS Lambda instance, which then pushes data—roughly 100,000 daily “click” events—into Amazon Kinesis, which in turn aggregates multiple streams of data in Amazon S3. SETTour then uses Amazon Athena to analyze its data in Amazon S3 to better understand website usage. The company also uses Amazon Elasticsearch Service to rapidly analyze website user logs and identify attempted access to website code or administration areas, which could indicate website hacks.

SETTour has been able to use the information in its big data environment to better understand user behavior. For instance, it has identified the increasing demand for videos on its website and the increasing amount of traffic stemming from mobile devices. “We now know that more than half of our visitors—some 60 percent—access our website using a mobile device,” says Lin. “This drives the need to showcase videos on our site because people prefer images to text on a small screen. With this kind of information, we can fine-tune our website’s design to better meet the needs of our customers.”

The company has improved the effectiveness of its marketing through its big data environment, which it uses to analyze the behavior of registered users on its website. Based on information such as the destinations most frequently viewed by users, SETTour tailors its email marketing messages. “Before we used to send out marketing emails in bulk, but now messages are personalized, based on behavior,” comments Lin. “Today, the number of marketing emails being opened by customers has increased from 6 to 40 percent. Furthermore, the number of those recipients that go on to buy holiday packages has risen from 1.1 to 5 percent, helping increase our revenues."