Amazon Live streams over 16 hours of Prime Day Deals over two days. While much of the production comes from the studio in New York, there are plenty of remote hosts and crew members that make it all happen.
The Challenge
For a streaming event of this profile, or any event of this size, you need backups. The best part of cloud production is the ability to have a distributed workflow. Switching and quality control operations can easily be shifted to a different location.
In the example signal flow below we are able to switch from the studio, or bypass the studio completely and switch to backup sources that are received directly by our cloud infrastructure.
The Signal Flow
While this isn't the exact signal flow we used for the Prime Day events, this is an example of a transmission and distribution flow that I would utilize for an event of this size. Primary and backup SRT encoders feed studio program to the cloud. Those encoders are received by our SRT contribution server (which includes auto-failover) If the primary goes down, the contribution server will automatically swap to the backup.