Tyson Vs Paul: A Lesson in Video Streaming Technologies

One of the most talked-about events from this past weekend was the premiere of a live boxing event streamed on Netflix, featuring the legendary Mike Tyson returning to the ring at 58 years old against social media star Jake Paul. Due to technical difficulties, what began as an intriguing must-watch event quickly turned into an internet meme. Throughout the night frustrated viewers experienced continuous frozen and buffering screens, disrupting the excitement and took away from the experience of Netflix's live broadcast.

As one of the pioneers in the streaming industry, Netflix revolutionized how we consume entertainment, shifting audiences away from traditional cable and satellite forms. It delivers high-quality video content to millions of users globally, leveraging sophisticated streaming technology to ensure smooth playback, scalability, and personalization. Netflix adeptly adapts to varying network conditions and device capabilities, reaffirming its role as an entertainment leader despite occasional setbacks.

Let’s uncover some of the differences in streaming live content vs on-demand, explore the unique technologies involved, what went wrong and how the industry can learn from it going forward.

How Protocols Differ Between Live Events and On-Demand Movies or Shows

Streaming Providers on-demand streaming and live event streaming use fundamentally different technologies and protocols due to the distinct challenges and requirements of each format. While movies and shows rely on pre-prepared content optimized for efficiency and quality, live events must handle real-time encoding and distribution, introducing complexities that require specialized streaming protocols. Here's a closer look at the differences:

1. Content Preparation

On-Demand

Before streaming providers content is streamed to viewers, it goes through an extensive preparation process:

• Video Encoding:
  Streaming providers use advanced video codecs like H.264 and H.265 (HEVC) to compress videos into smaller file sizes without significantly affecting quality.
  Each video is encoded into multiple resolutions (e.g., 480p, 720p, 1080p, 4K) and bitrates to cater to different network conditions.
• Audio and Subtitles:
  Streaming providers support multiple audio tracks and subtitle options. These are encoded separately to be dynamically combined during playback, reducing storage redundancy.
• Content Optimization:
  Using tools like Dynamic Optimizer, streaming providers fine-tune video files to ensure the best quality for a given bitrate, saving bandwidth without sacrificing user experience.
• DASH (Dynamic Adaptive Streaming over HTTP):
  Streaming providers primarily use DASH for its on-demand content. DASH breaks video into small chunks, allowing adaptive streaming based on a user’s bandwidth.
• HLS (HTTP Live Streaming):
  HLS is also supported, especially on Apple devices, and uses similar chunk-based adaptive streaming techniques.

Live Streaming

Live streaming requires real-time encoding and transcoding of video as the event happens. This introduces challenges in ensuring low latency while maintaining high quality.

RTMP (Real-Time Messaging Protocol):
Commonly used to deliver video from the source (e.g., event cameras) to streaming servers. RTMP is optimized for low-latency transport, critical for live streams.

Low-Latency HLS (LL-HLS) and CMAF:
To ensure real-time playback for live events, streaming providers likely use low-latency streaming protocols like LL-HLS combined with CMAF (Common Media Application Format) to reduce delay.

WebRTC:
WebRTC is a go-to for low-latency, real-time communication, potentially used for interactive live events.

Live CDN:
Streaming providers must adapt to handle the higher demands of live events, dynamically routing traffic to minimize disruptions and maintain a stable stream.

Dynamic Multicast Delivery: For live events, streaming providers must simultaneously deliver a single live feed to potentially millions of viewers, requiring multicast strategies or optimized unicast methods to avoid bottlenecks.

2. Content Delivery Network (CDN), Scalability and Bandwidth Usage

Streaming providers rely heavily on its proprietary networks to deliver videos efficiently.

On-Demand Content

• Regional Servers:
  Appliances are strategically placed in data centers and ISPs worldwide, allowing content to be stored closer to users.
  
• Asynchronous Playback:
  Since viewers watch content at their convenience, demand is spread out, reducing the likelihood of simultaneous traffic spikes.
• Caching:
  Popular titles are cached locally within these servers to reduce latency and minimize the load on the broader internet. For example, if a popular show has a new season, local servers are preloaded with the content before its release.
• Scalability:
  The distributed nature of networks ensures streaming providers can handle millions of simultaneous viewers without overwhelming a single server.

Live Events

• Sudden Traffic Spikes:
  Live events generate simultaneous demand from millions of viewers, creating immense pressure on servers. Streaming providers must provision additional resources and use protocols optimized for real-time delivery to handle these spikes.
• Dynamic Resource Allocation:
  Unlike on-demand content, live streaming requires streaming providers to dynamically allocate server resources to manage unpredictable traffic patterns.

3. Adaptive Streaming and Latency Management

On-Demand Content

Streaming providers employ Adaptive Bitrate Streaming (ABR) to deliver content seamlessly, even under fluctuating network conditions.

• How It Works:
  • The player on your device continuously monitors your internet speed.
  • Depending on available bandwidth, Netflix adjusts the video quality in real-time, switching between resolutions (e.g., from 1080p to 720p) to prevent buffering.
• Buffer Management:
  Streaming providers preload small chunks of video in advance, ensuring uninterrupted playback even if there are minor network interruptions.

Live Events

• Real-Time Constraints:
  Live events require ultra-low latency (often 5 seconds or less) to keep viewers in sync with the action. Protocols like LL-HLS and WebRTC are designed to prioritize real-time delivery, but they come with higher risks of buffering or quality drops if network conditions fluctuate.

4. Error Handling and User Experience, User Authentication 

Streaming providers backend systems handle user authentication and personalization, enhancing the overall experience:

On-Demand Content

• Pre-Optimized Delivery:
  Since on-demand content is pre-encoded and cached, error rates are low. Streaming providers algorithms monitor streaming quality and adjust bitrate or resolution in real time to prevent buffering.
• Authentication:
  When you log in, streaming providers verify your subscription status and grants access to its library. Geolocation services ensure you see the correct content catalog for your region.
• Recommendations:
  Streaming providers use machine learning algorithms to analyze viewing history, ratings, and preferences to recommend content tailored to each user.

Live Events

• Limited Recovery Options:
  For live events, interruptions or errors (e.g., buffering or server crashes) are more noticeable and harder to recover from since there’s no pre-buffered content. Streaming providers must rely on robust failover mechanisms to minimize disruptions.
• Viewer Synchronization:
  Ensuring that all viewers experience the live event in sync is a unique challenge for live streams, particularly when latency and buffering occur.

Key Technologies Driving Netflix Streaming

1. Video Codecs: H.264, H.265 (HEVC), AV1 (for even more efficient compression).
2. Protocols: DASH, HLS, and HTTPS for secure and efficient data transmission.
3. CDN: Open Connect for low-latency, high-speed delivery.
4. Encryption: HTTPS, DRM, and secure token authentication to protect content.
5. Algorithms: Machine learning for recommendations and adaptive streaming.

The Complexities of Live Streaming

Netflix’s struggles are not unique. Streaming live content is inherently more complex than delivering pre-recorded videos. Here’s why:

• Real-Time Encoding: Live streams must encode video and audio in real time, often introducing delays if the hardware or software falls short.
• Scalability: Predicting and managing traffic spikes is challenging, especially for highly anticipated events.
• Latency: Low-latency streams are essential for live events but require cutting-edge protocols and infrastructure.
• Global Distribution: Streaming live content across the globe means accommodating different internet speeds, device types, and regional server capabilities.

Lessons for the Industry

Netflix’s live event serves as a cautionary tale for the streaming industry. To avoid similar pitfalls, companies need to:

1. Invest in Scalable Infrastructure
   Platforms must overprepare for traffic spikes, using distributed cloud systems and robust CDNs to ensure stability.
2. Enhance Real-Time Monitoring
   Early detection of technical issues can minimize disruptions. Real-time monitoring tools should be used to detect bottlenecks or errors.
3. Test Under Load
   Simulating large-scale traffic during test runs can uncover weaknesses in the system before going live.
4. Communicate with Users
   If issues arise, platforms should communicate promptly and transparently with their audience to manage expectations.

Key Differences in Summary

While Netflix’s on-demand content streaming is a highly optimized and largely automated process, live events introduce complexities that require different protocols, real-time processing, and robust infrastructure to handle unpredictability. The challenges of latency, synchronization, and scalability highlight why live streaming is fundamentally different from the familiar world of movies and shows.

Netflix’s experience with live events, such as its recent foray into live specials, shows that even the most advanced streaming platform must adapt and innovate continuously to meet the unique demands of live content.

IVS Streamed Video Content

With the right technology and support from IVS, your institution can unlock the full potential of audio/video observation, recording, and streaming for training, skill development, research and evidence based applications.

To learn more about our unique expertise in streaming video content for your specific needs, contact our team today or request a free demo.