Active Packet Flow Disturbance for Classification Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing DPI solutions in telecommunications networks are passive and limited in classifying IP flows, struggling to handle the diverse characteristics of modern services, leading to inefficient packet classification and potential misclassification.
Innovation Solution
An apparatus and method that employs an active filter to disturb packet flows, allowing for more detailed analysis by monitoring reactions to classification attempts, using passive classifiers initially and adapting them based on observed reactions to improve classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If passive DPI classification is used, then device complexity is reduced, but measurement precision and classification reliability deteriorate
Solution Approach 1:
The patent implements feedback by monitoring reactions to disturbances in packet flows and using this information to adapt classification rules. The system observes how flows respond to active disturbances (such as bandwidth changes or packet delays) and adjusts classification based on these reactions, thereby improving classification accuracy without requiring overly complex passive analysis of all packet characteristics.
Solution Approach 2:
The system changes parameters of packet flows by actively introducing disturbances (modifying bandwidth, delay, or packet loss characteristics) to elicit reactions from the flows. These parameter changes serve as probes to identify flow characteristics that would be difficult to detect through passive inspection alone, thereby improving classification precision while maintaining manageable system complexity.
2Measurement precision
If active disturbance is applied to packet flows, then measurement precision improves, but loss of time increases due to additional monitoring and analysis
Solution Approach 1:
The patent applies periodic action by introducing disturbances at intervals rather than continuously monitoring all packet flows. The active filter periodically disturbs suspected flows and monitors reactions, allowing the system to achieve high measurement precision for critical flows while minimizing the overall time loss across the entire network traffic.
Solution Approach 2:
The system applies partial action by focusing active disturbance and monitoring only on packet flows that are suspected of being difficult to classify or that have high priority. Rather than applying continuous active analysis to all flows, the system concentrates resources on specific flows where improved classification accuracy is most needed, thereby reducing overall time loss while maintaining high measurement precision for critical cases.
Data Source
AI summary
Apparatus for performing packet classification of data packets belonging to a packet flow through a packet data network. The apparatus comprises an ingress interface for receiving packets of said packet flow, and an active filter for disturbing one or more of said packets or for otherwise disturbing a characteristic of said flow. An egress interface is provided for sending the packets including a disturbance towards a destination. A monitor is also provided for monitoring said packet flow and/or one or more associated packet flows received by the apparatus to detect subsequent reactions in the flow/s to the disturbance, while a flow classifier is provided for attempting to classify the flow into one of a set of defined classes according to a detected reaction.