Collaborative Traffic Classification via Aggregator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network traffic classification methods are computationally intensive and inefficient due to redundant analysis by multiple traffic classifiers, which can be addressed by implementing a collaborative approach that shares classification information and abstracted characteristics among classifiers.
Innovation Solution
A framework that facilitates collaboration between traffic classifiers through a central classification aggregator, allowing them to share and utilize mappings of abstracted characteristics to instantly classify communication flows, reducing redundant processing and improving scalability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple traffic classifiers independently classify network traffic, then each classifier can operate autonomously, but redundant analysis is performed and classification efficiency decreases
Solution Approach 1:
The patent merges the classification knowledge of multiple independent traffic classifiers into a single collaborative system. Classifiers share their classification results and abstracted characteristics through a common infrastructure, allowing them to leverage each other's analysis work. This combining approach eliminates redundant processing while maintaining autonomous operation of individual classifiers.
Solution Approach 2:
The patent implements preliminary action by having traffic classifiers generate and share abstracted characteristics (mappings) of communication flows in advance. When a classifier encounters a flow it cannot independently classify, it can query the shared database of abstracted characteristics created by other classifiers, avoiding the need to perform full analysis from scratch.
2Productivity
If traffic classifiers share classification information through a collaborative approach, then classification efficiency improves, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary component that manages the sharing of classification information between traffic classifiers. This intermediary infrastructure handles the complexity of coordination, data storage, and information retrieval, allowing individual classifiers to benefit from collaboration without directly managing the complex interactions between multiple classifiers.
3Productivity
If a large quantity of network traffic is classified instantaneously, then network control and efficiency are achieved, but computational intensity increases
Solution Approach 1:
The patent uses copying by creating abstracted representations (mappings) of communication flows that capture essential classification characteristics. Instead of performing full analysis on every traffic flow, classifiers copy and reuse these abstracted characteristics from the shared database, dramatically reducing computational intensity while maintaining classification accuracy and throughput.
Data Source
AI summary
Herein described is a collection of traffic classifiers communicatively coupled to a classification aggregator. Traffic classifiers may use conventional techniques to classify network traffic by application name, and thereafter may construct mappings that are used to more efficiently classify future network traffic. Mappings may associate one or more characteristics of a communication flow with an application name. In a collaborative approach, these mappings are shared among the traffic classifiers by means of the classification aggregator so that one traffic classifier can leverage the intelligence (e.g., mappings) formulated by another traffic classifier.