Unified Event Stream for Real-Time CDN Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content distribution and optimization systems face challenges such as breakage of single source of truth practices for analytics reporting, discrepancies in event models and metrics, requiring extensive maintenance and resource overhead, and inadequate real-time monitoring due to separate alarming and monitoring solutions, leading to delayed operational reporting and cumbersome debugging.
Innovation Solution
A content distribution and optimization system utilizing a publisher/subscriber model for event ingestion, enabling batch processing and real-time monitoring, with a heartbeat mechanism to derive new metrics and multiple use cases, and selecting suitable CDNs based on these metrics for efficient video quality of service delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If third-party tools with separate event streams are used for different use cases, then specific analytical requirements are met, but system complexity and maintenance overhead increase
Solution Approach 1:
The patent implements a universal event stream that serves multiple analytical use cases simultaneously. Instead of maintaining separate event streams for quality measurement, content reporting, and playback tracking, the system creates a single event stream that contains all necessary properties and dimensions. This universal event stream can be consumed by different analytical tools and services, reducing system complexity while maintaining versatility.
Solution Approach 2:
The patent merges multiple separate event streams into a single unified event stream. By combining quality measurement events, content reporting events, and playback tracking events into one stream with a standardized schema, the system eliminates the need to manage multiple separate data pipelines, reducing maintenance overhead and complexity.
2Reliability
If separate alarming and monitoring solutions are implemented, then specific monitoring requirements are met, but resource overhead and complexity increase
Solution Approach 1:
The patent creates a unified monitoring and alarming system that handles multiple monitoring requirements through a single solution. The event stream includes dimensions and properties that support both alarming and monitoring functions, eliminating the need for separate solutions and reducing maintenance overhead while maintaining comprehensive monitoring capability.
3Quantity of substance
If session level aggregate data is ingested daily from third-party tools, then reporting metrics are obtained, but real-time monitoring capability is lost
Solution Approach 1:
The patent implements real-time event ingestion and processing capabilities that prepare and make data available immediately when events occur, rather than waiting for daily aggregation. This allows the system to provide both immediate real-time monitoring and the necessary aggregate data for reporting metrics without the 24-hour delay associated with daily data ingestion.
Solution Approach 2:
The patent establishes continuous event streaming and processing that maintains uninterrupted data flow from client devices to analytical services. This continuous action ensures that monitoring data is always current and available in real-time, eliminating gaps and delays associated with periodic daily data ingestion.
Data Source
AI summary
Provided is a content distribution and optimization system that ingests raw event data from a client computing device at an entry point of a data pipeline service in accordance with a defined schema. Each payload of the raw event data comprises a first set of dimensional properties provided by the client computing device and/or a second set of dimensional properties added by the processor at the entry point. The raw event data is transmitted to a message bus pipeline for enrichment. A distinct use case is derived for each data consumer at a same time instant based on the enriched raw event data comprising same base event metrics associated with the base event. One or more payloads of the raw event data are transmitted to a stream-based messaging bus as raw video events. New metrics are derived based on raw video events for network selection and centralized alarming and reporting.