System and method for classifying and processing voice over IP traffic
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- SANDVINE CORP
- Filing Date
- 2021-09-29
- Publication Date
- 2026-05-21
AI Technical Summary
Existing systems struggle to accurately classify and process encrypted Voice over Internet Protocol (VoIP) traffic, limiting network operators' ability to manage congestion, ensure quality of experience, and perform billing and charging services, as metadata such as IP addresses and server names are often unavailable in encrypted traffic.
A system and method that collects traffic flow data, determines VoIP sessions through parameter correlation, and generates Internet Protocol Detail Records (IPDRs) or Call Detail Records (CDRs) by integrating raw data, using machine learning and heuristics to analyze encrypted VoIP traffic, even when IP sessions originate from different jurisdictions.
Enables accurate identification of VoIP sessions, providing detailed call records that include both parties' details, locations, and call attributes, enhancing network management and supporting law enforcement by correlating and visualizing data to generate useful intelligence.
Smart Images

Figure 0007863567000001 
Figure 0007863567000002 
Figure 0007863567000003