in

AI in Video Streaming

Conviva's AI-Driven Video Analytics

Conviva's video analytics platform harnesses the power of AI to deliver real-time insights, capturing data through proprietary sensors embedded in video players. This sophisticated system processes data from an impressive 500 million unique viewers and 180 billion streams annually, identifying playback issues to enhance user experience.

The platform's key components include:

  • Experience Insights: Monitors and diagnoses service delivery issues using AI-driven alerts, reducing the need for large operational teams dedicated to troubleshooting.
  • Ad Insights: Tracks ad performance and viewer engagement.
  • Social Insights: Provides cross-platform analysis of content performance on social media and streaming platforms.

Conviva developed a unique framework combining a time-state processor and a time-series database to handle complex queries over long sessions. This structure facilitates powerful stateful calculations and multidimensional analytics.

In collaboration with Datadog, Conviva extends its monitoring capabilities, offering users a comprehensive view of their video supply chain. This integration aids in troubleshooting issues across network components and evaluating factors like unauthorized user access.

A visual representation of AI-powered video analytics with data streams flowing from multiple devices

Integration with Datadog

The integration between Conviva and Datadog enhances video performance monitoring by merging Conviva's video analytics with Datadog's infrastructure telemetry capabilities. This collaboration enables clients to correlate playback activity with real-time infrastructure metrics, offering comprehensive and actionable insights.

Key benefits of the integration include:

  • Operational teams can monitor video performance and cross-reference it with network and infrastructure health metrics.
  • Datadog's advanced alerting and visualization tools complement Conviva's AI-driven insights.
  • Businesses can define specific MetricLenses for customized monitoring and alerting suited to their operational priorities.

The integration is particularly useful for detecting unauthorized usage or security breaches by comparing data sets from Conviva's playback metrics and Datadog's infrastructure metrics.

A split-screen view showing Conviva's video analytics alongside Datadog's infrastructure monitoring

Quality of Experience Metrics

Monitoring quality of experience (QoE) metrics like video start-up times and playback failures is crucial in the competitive streaming landscape. These metrics indicate overall user experience and impact viewer retention rates.

Video start-up time is a key first impression for users. Delays can prompt viewers to exit the platform or seek alternative content sources. Playback failures, such as buffering or sudden stops, disrupt viewer engagement and diminish perceived service reliability.

Conviva's platform provides granular insights into these QoE metrics through AI-driven analytics. This enables businesses to proactively manage start-up delays and playback anomalies through real-time data collection and intelligent alerts.

The integration with Datadog's infrastructure monitoring capabilities offers a comprehensive view of how QoE metrics intersect with underlying infrastructure performance. This allows businesses to correlate video playback data with network performance insights, facilitating prompt issue resolution.

Conviva's Ecosystem Module

Conviva's Ecosystem Module streamlines the detection and resolution of streaming issues, reducing operational costs and ensuring superior viewer experience. The module embeds proprietary sensors within video players to collect real-time data on various aspects of the streaming experience.

Key features of the Ecosystem Module:

  • AI-driven alerts identify root causes of streaming issues quickly, often without human intervention.
  • Continuous learning from ongoing data inputs, refining detection algorithms to adapt to new types of streaming disruptions or patterns.
  • Automation of common streaming issue resolutions, significantly reducing operational expenses.

By leveraging this module, companies can maintain optimal streaming standards with minimal oversight, allowing technical teams to focus on strategic initiatives.

A network of interconnected devices optimized by Conviva's Ecosystem Module

AI Alerts and Real-Time Diagnostics

Conviva's AI Alerts system analyzes real-time data collected from over 2.5 billion unique sensors embedded within video players. These sensors gather metrics on viewer engagement, playback performance, and Quality of Experience (QoE), enabling the platform to build a comprehensive Video Graph.

When anomalies are detected, the AI Alerts system:

  1. Triggers an automatic investigation
  2. Compares real-time QoE and engagement metrics against historic data
  3. Provides specific diagnostic data with each alert, including affected stream characteristics and potential causes

This proactive diagnostic capability enables video publishers to address problems quickly, often mitigating disruptions before they become noticeable to viewers.

By automating much of the detection and resolution process, Conviva's AI Alerts reduce the need for large operational teams traditionally required for monitoring and troubleshooting. This allows teams to focus on strategic initiatives aimed at business growth and innovation.

An AI system analyzing multiple video streams and generating real-time alerts

Conviva's integration of AI into video analytics offers an efficient approach to maintaining high-quality streaming service. This innovation enhances viewer satisfaction and strengthens the competitive position of service providers in the digital streaming industry.

  1. Zhang H, Ganjam A, Kelkar V. Conviva's Video AI Platform: Elevating OTT Businesses with Data-Driven Intelligence. IEEE Trans Network Serv Manag. 2021;18(2):1-12.
  2. Stoica I, Ganjam A. Real-Time Data Processing Challenges in Internet Video Streaming. Commun ACM. 2020;63(5):72-81.
Sam, the author

Written by Sam Camda

Leave a Reply

Your email address will not be published. Required fields are marked *

AI and IP Rights

Exploring the Latest in AI: Tools, Trends, and Practical Applications