Broadcasting live events such as a Football Championship on Free Ad-Supported Streaming TV (FAST) channels presents unique challenges, particularly due to the vastly increased number of concurrent users. Here's our solution.
Broadcasting live events such as a Football Championship on Free Ad-Supported Streaming TV (FAST) channels presents unique challenges, particularly due to the vastly increased number of concurrent users. Unlike regular programming, live events can attract audiences that are 10 to 100 times larger, or even more, than usual. This spike in viewership necessitates a robust and efficient strategy for delivering personalized ad content without compromising the viewing experience.
One major challenge is the requirement for personalized ad breaks for all viewers simultaneously, as live events do not allow for time-shifted viewing. Simply relying on horizontal auto-scaling to manage this peak demand is not a viable solution. Auto-scaling can be slow to respond to sudden surges in user numbers and may not efficiently handle the specific demands of delivering personalized ads in real-time.
To address this, a prefetch mechanism that prepares ad breaks in advance is a far more effective solution. This process involves several detailed steps:
Firstly, integrating RapidStream into the broadcast workflow requires ingesting the FAST schedule, which details the ad break time slots. As these times approach, RapidStream prefetches the active user sessions associated with the FAST channel ID based on this schedule. This prefetching is crucial as it allows the system to manage the high volume of concurrent user requests more efficiently.
The list of active user sessions is then used to generate personalized ad breaks. This is achieved by querying ad servers according to predefined ad rules and inventory share agreements. The personalized ad breaks are created by considering three key data contexts: the Electronic Program Guide (EPG) context, which includes metadata about the current event; the streaming context, which comprises metadata about the stream and active user sessions; and the user context, which analyzes user preferences based on their viewing behavior.
The ad server responses are stored in a file bucket, each identified by a unique user session identifier. When it’s time to insert ads into the stream, the Server-Side Ad Insertion (SSAI) service requests the ad break data using this unique session identifier, expecting a VAST document from RapidStream. RapidStream then redirects this request to the file bucket, allowing the ads to be seamlessly stitched into the stream.
In addition to managing ad breaks efficiently, RapidStream also enhances analytics capabilities. By providing a prefetchId
to all processes, it allows analytics to be clustered, enabling the analysis of Key Performance Indicators (KPIs) for all users within the same ad break. This data is invaluable for understanding viewer engagement and ad performance during live events.
In conclusion, broadcasting a live event like a Football Championship on a FAST channel demands innovative solutions to handle the surge in concurrent users and the need for synchronized, personalized ad content. By employing a prefetch mechanism and integrating RapidStream’s advanced features, broadcasters can ensure a seamless viewing experience, delivering ads effectively without disrupting the event’s live nature.