Streaming Session Simulation for Bitrate Model Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Evaluating bitrate selection algorithms in adaptive bitrate streaming is computationally expensive and requires significant memory usage, necessitating large amounts of content streaming and data gathering to assess improvements in streaming experiences.
Innovation Solution
A media device simulation system aggregates streaming session data, determines segment parameters, and performs simulations using different bitrate selection models to evaluate and compare their performance without actual content streaming, thereby selecting the most effective model for improved streaming experiences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If actual content streaming is performed to evaluate bitrate selection algorithms, then evaluation accuracy is improved, but computational cost and memory usage increase significantly
Solution Approach 1:
The patent creates a simulation environment that copies the essential characteristics of actual streaming sessions without requiring real content streaming. The system generates synthetic streaming sessions that replicate network conditions, device constraints, and user behavior patterns, allowing evaluation of bitrate selection algorithms through simulation rather than actual content delivery. This approach maintains evaluation accuracy while dramatically reducing computational resources and energy consumption.
Solution Approach 2:
The patent introduces a simulation system as an intermediary between the bitrate selection algorithms and actual streaming content. Instead of directly evaluating algorithms through real content streaming, the simulation system acts as a mediator that translates algorithm performance into measurable metrics. This intermediary layer enables accurate evaluation without the computational burden of actual content streaming operations.
2Measurement precision
If large amounts of content are streamed for evaluation, then evaluation accuracy is improved, but memory usage and data gathering requirements increase
Solution Approach 1:
The simulation system generates synthetic streaming session data that copies the statistical characteristics of real streaming patterns without requiring actual content streams. By modeling network conditions, device constraints, and user behavior through probability distributions and statistical parameters, the system creates representative evaluation datasets that maintain accuracy while using minimal storage space and data volume.
Solution Approach 2:
The patent transforms the evaluation approach by changing from concrete content data to abstract statistical parameters. Instead of storing and processing actual streaming content, the system uses parameterized models that represent network conditions, device capabilities, and user behavior. This parameter transformation enables accurate evaluation with significantly reduced data quantity requirements.
3Reliability
If actual streaming sessions are gathered for evaluation, then evaluation reliability is improved, but time required for data collection increases
Solution Approach 1:
The simulation system performs preliminary generation of streaming session data before actual evaluation begins. By pre-generating diverse streaming scenarios, network conditions, and device states through the simulation engine, the system prepares comprehensive evaluation datasets in advance. This preliminary action enables rapid evaluation of multiple algorithms without time-consuming real-world data collection, while maintaining reliability through statistically representative scenarios.
Data Source
AI summary
Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for evaluating bitrate selection models. An example embodiment operates by aggregating data associated with a data streaming session comprising a plurality of segments. The embodiment then determines, for each segment of the plurality of segments, segment data comprising streaming parameters by analyzing the aggregated data. The embodiment then generates a simulation input based on the segment data for each segment. The embodiment then performs a first simulation, based at least on the segment data for each segment and a first bitrate selection model, to simulate streaming each segment. The embodiment then generates a first simulation result for each segment. The embodiment then aggregates each first simulation result to provide a first aggregate simulation result of the data streaming session.