iFit Case Study

Taiwan-based iFit started out as an online community for fitness and weight loss. The business has evolved since launching in 2012, becoming an e-commerce platform for health and wellness products as well as a traditional brick-and-mortar retailer with stores across Taiwan. In 2014, iFit received about NT$15 million (US$488,000) in venture capital funding to help drive growth across East Asia, particularly China. Revenues and the size of the iFit community have doubled due to the success of its online marketing campaigns.

iFit runs most online campaigns on platforms such as LINE and Facebook to highlight new products or special offers. The number of visitors on these platforms can suddenly increase by 50 times while campaigns are running. In order to obtain the best browsing experience for customers and close as many deals as possible, iFit’s IT infrastructure must either scale up to support these spikes or risk reduced performance.

From the time it launched, iFit ran its IT on a cloud platform. However, as the business grew, limitations of that platform surfaced and it began to require a lot of management. “There were a lot of repetitive jobs that took up our time,” says Leonard Chuang, chief information officer of iFit. “For example, scaling web servers to support a campaign meant installing the OS and manually setting up the database each time. These web servers and databases then had to be taken down at the end of the campaign, which took even more time.” In essence, Chuang wanted a cloud service that was easier to scale up and down according to the level of campaign activity. He also aimed to cut back on basic management time so IT staff could focus on developing site features.

iFit decided to switch from its existing cloud-service provider to Amazon Web Services (AWS) in April 2014. “We had been researching AWS for a while and saw that we could automate enough processes to bring down our management costs,” says Chuang. The company completed the migration to AWS in less than one week. The IT team switched both the web servers and database over to AWS using a file transfer protocol and AWS Mobile SDK for Java.

Incoming website traffic now hits the domain name service provided by Amazon Route 53 and passes through Elastic Load Balancing services before being parsed by web and application services running on Amazon Elastic Compute Cloud (Amazon EC2) instances. At this point, product images are rendered from Amazon Simple Storage Service (Amazon S3) using the Amazon CloudFront content delivery network to reduce latency. iFit relies on Amazon Relational Database Service (Amazon RDS) to process transactions. Once a transaction has been completed, the company takes advantage of Amazon Simple Email Service (Amazon SES) to send the shopper an invoice confirming the order.

iFit is now running infrastructure for its website and e-commerce platform on AWS. Chuang has achieved his main goal of simplifying infrastructure management. “We have reduced management time by up to 80 percent with AWS compared to our previous cloud platform,” says Chuang. As a result, the company has been able to redeploy three IT staff members for development tasks, enabling the company to double the speed of expanding website features.

Because the AWS platform takes much less time to administer, management overhead has also been significantly lowered. “Effectively, we can scale our infrastructure to support business growth at one-third the cost of our previous environment,” says Chuang.

In addition to being simpler to use, using AWS also provides iFit with more flexibility. IFit can now scale their infrastructure in half a day, compared to one or two days in its previous environment, making it easier for the website to handle sudden increases in visitor numbers during campaigns. “Thanks to this flexibility, we can also scale down our services when demand falls,” says Chuang.

Chuang is also pleased with the performance of the iFit website on the AWS platform. While the backend IT infrastructure is located in the AWS US West (Oregon) Region, front-end IT services run from the AWS Edge Network Location in Taiwan. “Latency is very low and read/write times are very fast with AWS,” Chuang says.

Looking ahead, Chuang is excited to introduce some form of machine learning to the iFit platform to help tailor website visits for users. According to Chuang, AWS offers the capacity and compute performance to make this happen down the line. “With AWS, I see great potential for enhancing our own internal efficiencies while improving the experience of our site visitors.”