Scaled Web Service

“Rooster Park’s engineers designed and built the service we needed now, and one we’re going to be able to use for years to come.”

 

Challenge

The world’s largest sports and entertainment company wanted to show relevant news, scores, and stories to any user who visited their website or mobile application. They have a range of customers¬†– people who visit multiple times each day, obsessively following every bit of news for their favorite team, and people who visit occasionally to see what’s going on locally and nationally.

This service needed to show relevant data to both kinds of visitors (whether they were logged in or not), learn from what customers did and improve over time, and respond to thousands of simultaneous worldwide transactions. It also had to integrate with a variety of customer-facing products on different ship schedules, and be ready for a major redesign.

 

Solution

The Rooster Park team built a system that could do both explicit personalization¬†(show users Seahawks information because they requested it) and¬†implicit personalization (choose information based on a visitor’s location, usage pattern, visit patterns from other customers, etc.), and could respond to service requests quickly. This worked for both casual readers and fantasy sports players.

The team built RESTful services in Java that could handle >15,000 transactions/second, building core data infrastructure (MongoDB/Hadoop) and a reporting infrastructure (MongoDB/Hadoop/Hive/Pig). They worked with teams based out of Seattle and from the East Coast, and introduced agile development methodologies to parts of the organization.