AWS Case Study: Ooyala
Ooyala is an international video technology platform and solutions company that helps ESPN, Bloomberg, and other companies take video content from traditional TV and provide it to users online. Customers can leverage Ooyala’s video analytics to make informed decisions around creating, distributing, and monetizing content. The company needs capacity to store content that grows continuously, and has to be able respond to spikes in traffic when unexpected news events occur. By running on Amazon Web Services (AWS), Ooyala is able to store 1.5 petabytes of data and can scale its infrastructure to almost 5 times its normal size.
The company was an early user of Amazon Web Services (AWS) and the winner of the first AWS startup challenge. Today, AWS is leveraged as part of Ooyala’s elastic technology stack that allows it to quickly access additional capacity if needed for transcoding/encoding, content management and data processing.
Why Amazon Web Services
Ooyala leverages AWS Reserved and On-Demand instances. Ooyala’s customers use a web application called Backlot and fully extensible APIs to manage, analyze, monetize, and publish their content. The application and APIs make heavy use of Amazon EC2, which connects directly to Ooyala’s management servers. Amazon EC2 is ideal for Ooyala because it is able to quickly scale to manage increased load and to not invest in its own data center build-out ahead of demand. Ooyala improves ROI by purchasing Reserved Instances when they believe they will be used more than 50% of the time.
For encoding, Ooyala has a centralized job management system built on Resque. As new encoding jobs come in they are immediately added to a queue to be processed. Ooyala uses Amazon EC2 to launch instances to remove jobs from the queue and perform the encoding process based on the pre-defined encoding profile. Ooyala uses a combination of Amazon EC2 Spot, On-Demand and Reserved Instances to cover the baseline-predicted volume of video ingestion/encoding jobs. To minimize their costs, Ooyala bids on Spot Instances below the On-Demand rate but higher than the Amazon EC2 Reserved Instance hourly rate. To manage the potential that their Spot Instances may be interrupted, Ooyala’s encoding management software allows for jobs to fail at random.
Sean Knapp, Founder and Chief Technology Officer at Ooyala, observes, “When we started Ooyala, we were warned that we would spend most of our time in datacenters at 2:00 AM, making sure everything was working. With AWS, we have found ourselves asking why anyone would choose to do it themselves. AWS has enabled us to build, deploy, and scale our product in record time, raising the bar for rapid innovation.”
Ooyala leverages Amazon Simple Storage Service (Amazon S3) for storage and delivery of its vast library of videos. The company stores large amounts of video files easily within Amazon S3, benefitting from the service’s scalability and tools that help Ooyala secure its data.
Using standard REST interfaces from Ruby, C++, and Java programs, Ooyala was able to build a solution on AWS rapidly and easily, reducing development time and accelerating time to market. Knapp cites the speed of building and deploying on AWS as an advantage over competitors who are still using traditional hosting models.
Based on its positive experience with AWS, the Ooyala team is evaluating the use of additional AWS services like Amazon ElasticMap Reduce. Knapp advises others using AWS to do the same: “As AWS provides an increasing number of services, companies should take care to properly architect their solutions to intelligently leverage the appropriate capabilities of the broader offering.”
To learn more about how AWS can help your digital media needs, visit our Digital Media details page: http://aws.amazon.com/digital-media/.