Enriching Encrypted Traffic Analytics via Flow Data Mapping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Many networks lack the necessary hardware to fully exploit Encrypted Traffic Analytics (ETA) due to insufficient capabilities for extracting ETA telemetry data, resulting in incomplete benefits from this valuable information source.

Innovation Solution

A method is developed to generate a database that maps traffic flow information features to encrypted traffic analytics features, allowing for the determination of encrypted traffic analytics features for domains lacking this information using machine learning models constructed from domains with both types of data available, thereby enriching knowledge about network traffic.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If Encrypted Traffic Analytics (ETA) is deployed in networks, then valuable information about network traffic and devices can be extracted, but hardware capability requirements become too high for many networks

Engineering Contradiction:
ImproveETA information availabilityVSAvoidhardware capability requirements
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent creates a mapping model that copies ETA features from domains with ETA capability to domains without ETA capability. By using traffic flow information features as a proxy and establishing correspondence relationships through machine learning models, the system replicates ETA information for domains where direct ETA extraction is not available, thereby reducing hardware requirements while maintaining information availability.

Inventive Principle:
Principle #26Copying

2Adaptability or versatility

If machine learning models are trained with both traffic flow information and ETA information, then ETA features can be determined for domains with only traffic flow data, but data requirements and processing complexity increase

Engineering Contradiction:
ImproveETA feature determination capabilityVSAvoiddata requirements
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The patent extracts the essential mapping relationship between traffic flow information features and ETA features through machine learning models. By training on domains with both types of data and then applying the learned mapping to domains with only traffic flow data, the system extracts and transfers only the necessary feature correspondence information, reducing the actual data requirements during deployment while maintaining adaptability.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3718284B1Extending encrypted traffic analytics with traffic flow data
Publication Date: 2023.01.04 CISCO TECHNOLOGY INC
  • EP3718284B1 patent drawingFigure 1
  • EP3718284B1 patent drawingFigure 2
  • EP3718284B1 patent drawingFigure 3

AI summary

Techniques for enriching encrypted traffic analytics are presented, In one embodiment, a method includes obtaining telemetry data for one or more domains within a network. The telemetry data includes both encrypted traffic analytics information and traffic flow information associated with the network traffic. For each domain of the one or more domains, the method also includes generating a model comprising a mapping from a plurality of traffic flow information features to at least one encrypted traffic analytics feature. The method includes generating a database comprising generated models for each of the domains and obtaining telemetry data for a target domain that includes traffic flow information, but does not include encrypted traffic analytics information. At least one encrypted traffic analytics feature of the target domain is determined based on a plurality of traffic flow information features of the target domain using the database.