AWS Case Study: Fliptop
Fliptop, based in San Francisco and Taipei, is a marketing platform that helps brands convert their e-mail lists into social profiles. The primary function of Fliptop's initial product is social profile lookups. Given an e-mail address, they use over two dozen publicly available data sources and a sophisticated matching algorithm to determine social network memberships, demographics, and psychographics.
Amazon Elastic MapReduce has enabled Fliptop to scale their lookup capacity to over one million contacts per day. While it would have taken months to build a traditional data center of this size, it took only weeks with Amazon Web Services (AWS). Chiao says, "AWS took away our last excuse for not committing to the Hadoop platform. Amazon Elastic MapReduce with Spot Instances has made it easy to prototype and surprisingly cost-effective to scale, decreasing our data processing costs by over 50%."
Why Amazon Web Services
Outside of their domain name service (DNS) the Fliptop infrastructure runs entirely on AWS, including the following services:
- Amazon Elastic MapReduce
- Amazon Simple Storage Service (Amazon S3)
- Amazon Elastic Compute Cloud (Amazon EC2)
- Amazon Relational Database Service (Amazon RDS)
- Elastic Load Balancing
- Amazon CloudFront
- Amazon CloudWatch
- Amazon Simple Email Service (Amazon SES)
Fliptop's technology stack includes Ruby on Rails, Scala, and Java. Chiao adds, "We use a variety of open-source tools, like Apache Solr for search, Drools for business rules, and Hadoop (via Amazon Elastic MapReduce) for big-data crunching." The Fliptop team also uses the Amazon SDK extensively to automate their Elastic MapReduce jobs and for various operations tasks.
Chiao notes, "We used to maintain an SMTP server in another hosting platform, but have since switched over to Amazon SES due to simplicity of use and for the benefit of having our entire infrastructure in one place, with one monthly bill."
AWS allows the company to scale to hundreds of instances to process massive lookup jobs without the corresponding capital expenditure of a traditional infrastructure. "In addition," says Chiao, "the built-in redundancy, recovery, and monitoring features of services like Amazon RDS allow our development team to run our production infrastructure without any dedicated operations resources."
Chiao recommends that other developers plan their AWS infrastructure so services and data are available across multiple availability zones and regions: "While outages are rare, developers should take advantage of the AWS services that make it easy to avoid them when they occur."
The company is growing quickly and looking for talented big-data engineers and data scientists. Chiao comments, "Our customers loaded tens of millions of contacts into our system in just the first few weeks after launch. There's no way we could have handled that demand without AWS."
To find out more about how AWS can help you store and process big data, visit our Big Data details page: http://aws.amazon.com/big-data/.