BroadbandTV powers YouTube content creation, optimization and monetization

The obstacles they faced

Supporting scalable content optimization tools

Committed to using technology to provide services to its partners, BBTV built products using state of the art algorithms to improve viewership. In addition, the company built metrics for insights and optimized videos for consumption. BBTV needed to deliver these innovative products and features at scale with an infrastructure that could handle big data. Security and high availability were also big concerns for the business.

Designing a secure isolated deployment protected from DDoS attacks had to be a priority. Given the pace with which the company was moving forward, the need to automate the deployment was of an equal priority.

What we achieved together

The relationship between BBTV and Rackspace Technology continues today, focusing on areas like cost optimization, the introduction of various testing and sandbox environments to support the internal agile process and the growing number of products in development, best practices, and policies for access and management of AWS resources, operational management of the production systems, etc.

The deployment of the products coming online at the time was scripted to take advantage of Elastic Beanstalk and RDS, reducing the pain of deploying to a complex clustered environment and removing the most typical tasks for maintaining and operating the environment while making it easy to scale.

As BBTV’s products like VISO Catalyst were successfully rolled out to all BBTV partners, the need to process massive amounts of video and accelerate the onboarding process of new partners led to its deployment of more products that had to perform extremely fast.

Cost was a significant risk for the company and lots of design thoughts went into designing the deployment to take advantage of the AWS spot pricing and reduce the bandwidth for moving data in and out of the network.

Rackspace Technology and BBTV have been able to architect a highly de-coupled solution using spot instances and Amazon SQS workers to process around half a million YouTube videos in a few weeks for a fraction of the cost it would take to do that with one of the largest instances available if the product was run on it in the same way it was tested internally before the release.