Building Streaming Infrastructure That Scales: Because Viewers Won't Wait Until Tomorrow

Background

ProSiebenSat.1 Media SE, one of Europe’s largest broadcasters, faced the challenge of scaling streaming applications for international users. The company’s streaming application, Joyn, handles millions of requests across Austria, Switzerland, and Germany, requiring highly available and scalable systems with minimal cost.

The original architecture struggled with overloaded servers, database issues, and data inconsistencies, causing frequent system crashes during peak viewership. Through iterative improvements, and leveraging serverless technologies, the team significantly increased availability and scalability.

Why This Matters

Idealized cloud architectures often clash with cost realities and operational complexities. The example demonstrates that a team with limited initial experience can radically improve system performance and reliability, but requires a focus on continuous iteration, automation, and clear communication of trade-offs between availability, cost, and complexity – failures in streaming translate directly to lost viewers and revenue.

Key Insights

• Hub and Spoke pattern: Provides clear service boundaries and data consistency.
• Cell-based architecture: Reduces blast radius by splitting traffic across regions, user types, and platforms.
• Multi-layer caching: Reduced database load to under ten percent, enabling more cost-effective multi-region strategies.

Working Example

# Example of EventBridge Rule (Conceptual)
{
  "source": "com.example.streamingservice",
  "detail-type": "VideoViewed",
  "detail": {
    "video_id": "12345",
    "user_id": "67890",
    "timestamp": "2025-12-24T10:00:00Z"
  }
}

Practical Applications

• Use Case: ProSiebenSat.1 Media SE improved scalability and resilience of its Joyn streaming platform.
• Pitfall: Exposing internal service state on a company bus (e.g. Kafka) without proper infrastructure creates a recognized anti-pattern.

References:

• https://www.infoq.com/articles/evolution-backend-streaming-application/