Hierarchical Media Session Tracking via Nested Flow and Interaction Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional tools are inadequate for accurately analyzing modern media sessions due to complexities arising from parallel TCP flows and multiple server deliveries, making it difficult to track and analyze media streaming effectively.
Innovation Solution
A network device employs a hierarchical modeling scheme to identify, organize, and analyze media sessions by instantiating flow, interaction, subscriber, session, and stream models, which process transport flows, application-level interactions, and metadata to generate session-based analytics and metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional tools are used to analyze media sessions, then network traffic can be monitored using standard metrics, but accurate analysis of media session delivery quality cannot be achieved due to parallel TCP flows and multiple HTTP interactions
Solution Approach 1:
The patent segments the complex media session tracking problem into distinct manageable components: flow-level tracking (identifying parallel TCP flows), interaction-level tracking (identifying HTTP requests/responses within flows), and session-level aggregation (combining fragments into complete sessions). This segmentation allows each component to be handled with dedicated logic, making the overall complex task achievable.
Solution Approach 2:
The patent implements a nested hierarchical model where flow models contain interaction models, which in turn contain session models. Each nested level aggregates information from the level below it, allowing the system to manage complexity at multiple granularities simultaneously while maintaining accurate end-to-end session analysis.
2Measurement precision
If all HTTP interactions are searched and fragments are combined to complete media session analysis, then accurate session modeling is achieved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent performs preliminary actions by pre-establishing flow models and interaction models as the media sessions progress. Flow models are created when parallel TCP flows are detected, and interaction models are instantiated when HTTP requests are observed. This preliminary structuring allows fragments to be immediately categorized and aggregated into sessions without requiring a complete search through all HTTP interactions, significantly reducing processing time while maintaining accuracy.
3Reliability
If multiple servers are used to deliver media content with multiple flows and interactions per server, then delivery reliability and optimization are improved, but the complexity of tracking and analyzing individual sessions increases
Solution Approach 1:
The patent creates universal models that can handle multiple servers and delivery configurations through a single unified approach. Flow models, interaction models, and session models are designed to work consistently whether media comes from a single server or multiple servers, whether using single or parallel flows. This universality eliminates the need for server-specific tracking logic, reducing overall system complexity while maintaining the ability to analyze reliable multi-server deliveries.
Data Source
AI summary
In a networked system connecting a subscriber to a content source via one or more networks, a computer-implemented method includes instantiating a session model for a media session between the subscriber and the content source, the session model to generate session metadata from application-layer interactions between the subscriber and the content source. The method further includes instantiating an interaction model in response to detecting an application-layer interaction in a transport flow of the subscriber, the application-layer interaction comprising media data for the media session. The method also includes generating, at the interaction model, interaction metadata representative of the application-layer interaction, and processing at least one of the interaction metadata and the media data at the session model to generate session metadata responsive to matching the interaction model to the session model.