Local Device Agent for QoE-Aware Packet Flow Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in identifying and matching specific packet flows to applications in computer networks, particularly in ensuring Quality of Experience (QoE) by accurately prioritizing and treating network flows.
Innovation Solution
A local device agent maps packet flows with application sessions based on execution information captured by the operating system, determines adjustments according to application profiles, and applies these adjustments to packets for optimal processing and QoE enhancement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If DSCP marking is used for QoS treatment, then network flow prioritization is improved, but correlation with Quality of Experience (QoE) is worsened
Solution Approach 1:
The patent introduces an intermediary component (QoE measurement and feedback mechanism) that bridges the gap between DSCP marking and actual QoE. This intermediary captures real QoE metrics, compares them with expected values, and provides feedback to adjust marking strategies, thereby restoring the lost correlation between network treatment and user experience.
Solution Approach 2:
The patent implements a feedback loop where QoE measurements are continuously monitored, analyzed, and used to refine flow identification and marking decisions. This feedback mechanism enables the system to learn from actual QoE outcomes and improve its ability to correlate network prioritization with user experience over time.
2Measurement precision
If packet flow identification is made more granular, then QoE-aware processing is improved, but device complexity is worsened
Solution Approach 1:
The patent segments the packet flow identification process into multiple independent components: flow key extraction, application session matching, QoE metric collection, and marking decision generation. Each component handles a specific aspect of the problem, reducing overall system complexity while maintaining granular identification capabilities.
Solution Approach 2:
The patent adds a new dimension to flow identification by incorporating QoE metrics and application session context beyond traditional five-tuple matching. This dimensional expansion enables more precise QoE-aware processing without proportionally increasing complexity, as the additional dimensions are integrated through structured data models and algorithms.
Data Source
AI summary
In one embodiment, an agent executed by a device maps a packet flow of the device with a session of an application executed by the device based on execution information about the application that is captured by an operating system of the device. The agent determines, based in part on mapping the packet flow of the device with the session of the application, one or more adjustments for a packet of the packet flow according to an application profile. The agent forms an adjusted packet by applying the one or more adjustments to a packet of the packet flow, wherein the one or more adjustments indicate to a networking device how packets of the packet flow should be processed. The agent sends the adjusted packet as part of the packet flow to an external destination via a network.