Hudl Case Study

Hudl is a software provider that offers a video and analytics platform for coaches and athletes to quickly review game footage to improve team play. The web-based system serves youth, amateur, and professional teams in football, basketball, soccer, volleyball, and other popular sports. Hudl was named to the 2016 Most Innovative Companies list by Fast Company.

Hudl is experiencing significant business growth around the world, with an increasing number of sports teams using the company’s services to upload game video. “We are growing at about 30 percent every year, so we need the scalability to support and maintain that growth,” says Brian Kaiser, CTO at Hudl.

The company also strives to provide faster video upload speeds to its customers. “A lot of coaches were asking for us to enable faster video uploads and get them video sooner so they could analyze it more quickly,” says Kaiser. “That is a logistical challenge because we’re moving large video files over long distances. On a typical Friday night in high school football season, for example, we upload 39 hours of video every minute for encoding and processing. That requires the right technology on the back end.”

Hudl also sought a better analytical platform for internal data analysis. “We built our own data warehouse, but it was complicated and expensive, and it didn’t perform well,” Kaiser says. “We needed to address that issue because we needed to get faster and better project analytics.”

Having launched its platform on the Amazon Web Services (AWS) Cloud, Hudl realized it needed to expand its use of AWS services to meet its needs for scalability, faster uploads, and more effective data analysis.

The company had been running its platform on Amazon Elastic Compute Cloud (Amazon EC2) instances and Amazon Simple Storage Service (Amazon S3) buckets, which it decided to enhance by incorporating Auto Scaling. “Auto Scaling gives us an automated way to scale storage and compute capacity,” says Kaiser. “We run a microservices architecture, and each of our clusters is in an Auto Scaling group. That helps us because our load varies from week to week. Using Auto Scaling, we don’t need to manage the process of scaling down, so we avoid the cost of over-scaling.”

To meet its needs for faster video uploads, Hudl began using Amazon S3 Transfer Acceleration, a feature that enables faster data transfers to and from Amazon S3. Hudl also started taking advantage of Amazon Redshift as its data warehouse for internal data analysis. The organization also uses AWS to support an additional data analytics and predictive-learning platform, which customers use to analyze opponents’ tendencies and other game trends. Hudl also uses Amazon CloudFront as its global content delivery network (CDN) service for fast delivery of its video platform. Hudl uses Amazon ElastiCache for Redis to provide millions of coaches and sports analysts with near-real-time data feeds, so they are better equipped to drive their teams to victory.

Hudl can keep pace with its 30 percent annual growth by running its video platform on AWS. The company uses AWS to ingest and encode more than 39 hours of high-definition video every minute during sports seasons, for 4.5 million coaches and athletes on 130,000 teams worldwide. “We have more and more global customers every year, and AWS gives us the scalability and performance to support them all,” says Kaiser. “Using AWS, we can easily spin up 2,000 servers just for video encoding on any given Friday night during football season. We never run into capacity problems because scaling is controlled automatically with Auto Scaling groups.”

With Transfer Acceleration, Hudl has increased video upload speeds on its platform. “Using Amazon S3 Transfer Acceleration, we’ve seen at least a 20 percent increase in video upload and encoding speeds,” says Kaiser. “That means high school football coaches, for example, can do their video uploading and analysis faster and get home to their families sooner. So in addition to improving their teams, they can have more time to themselves. That’s something we take a lot of pride in as a company.”

Internal data analysis is now better and more cost effective for Hudl. “Amazon Redshift has been amazing for us in terms of performance consistency and usability,” Kaiser says. “When we tried building our own data warehouse, the performance was highly volatile, but that has gone away with Amazon Redshift. And as far as usability, more than 50 percent of our company has written Redshift queries and run them against our Redshift cluster. That’s a testament to how well it performs and how fast it delivers actionable data back to users.”

The company is also greatly reducing the cost of running internal data and predictive analytics. “Our cost per compute or storage unit is down significantly year over year, because of the capabilities of AWS,” says Kaiser. “And we plan to expand the number of AWS services we use in the near future. There’s no question we’ll be able to optimize the performance and costs of our entire platform even more.