ML Network Session Correlation for Malware Beacon Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cloud computing platforms face challenges in detecting malware activity, particularly through network communication channels like command-and-control channels, as malware often employs obfuscated initialization routines and beaconing mechanisms that mimic benign periodic activity, making detection difficult.

Innovation Solution

A networked computing system employs a malware detection system that analyzes network and process session records using machine learning models to identify potential malware activity, generates alerts, and performs mitigation operations such as terminating processes or blocking network communication.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional detection methods are used to identify malware, then false positives increase due to inability to distinguish from benign periodic activity, but detection accuracy decreases

Engineering Contradiction:
Improvedetection accuracyVSAvoidfalse positive rate
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The detection system segments the analysis into multiple independent components: network traffic pattern analysis, beaconing behavior detection, correlation with known malware signatures, and benign activity profiling. Each segment operates independently to evaluate specific aspects of the traffic, allowing the system to achieve high detection accuracy while maintaining reliability by cross-validating results across multiple segments before generating alerts.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If comprehensive analysis of all network traffic is performed, then detection accuracy improves, but system complexity and processing time increase

Engineering Contradiction:
Improvedetection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system applies local quality by implementing specialized detection modules for specific traffic characteristics: one module analyzes periodic beaconing patterns, another examines command-and-control communication signatures, and a third profiles benign organizational traffic. Each module is optimized for its specific function with appropriate complexity level, allowing comprehensive analysis without requiring the entire system to be uniformly complex. This modular approach maintains detection accuracy while managing system complexity.

Inventive Principle:
Principle #3Local quality

3Speed

If real-time analysis of network traffic is implemented, then response time improves, but computational resource consumption increases

Engineering Contradiction:
Improveresponse timeVSAvoidcomputational resource consumption
Core Design Contradiction:
SpeedVSUse of energy by moving object

Solution Approach 1:

The system implements periodic action by analyzing network traffic at strategically determined intervals rather than continuously processing every packet in real-time. The beaconing detection module periodically samples traffic patterns at intervals matching typical malware beaconing periods, while maintaining the capability to trigger immediate analysis when anomaly thresholds are exceeded. This approach achieves effective real-time response for malicious activities while significantly reducing computational resource consumption during normal operation.

Inventive Principle:
Principle #19Periodic action

Data Source

PatentEP4533302B1Malware activity detection for networked computing systems
Publication Date: 2025.12.17 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4533302B1 patent drawingFigure 1
  • EP4533302B1 patent drawingFigure 2~3
  • EP4533302B1 patent drawingFigure 4

AI summary

Malware activity detection for networked computing systems is described. A network session record is provided to a machine learning (ML) model configured to generate an indication of whether the provided network session record evidences malware activity. The network session record indicates network traffic activity in a time period. Responsive to an indication by the ML model, correlation scores are calculated by, for each process session record in a process session record set, calculating a correlation score indicative of a correlation between the provided network session record and the process session record. Each process session record in the process session record set corresponds to a process executed by a computing device in the time period. A determination that a correlation score indicates a corresponding process session record is indicative of the evidenced malware activity is made. Responsive to the determination, a malware activity alert is generated.