Telecom Fraud Detection via Shared Transaction Registry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for detecting fraud in international telecommunications lack end-to-end visibility, leading to various types of fraud such as call hijacking, Simbox fraud, and CLI manipulation, resulting in substantial losses for operators and carriers, and existing solutions are limited in detecting a wide range of fraud types reliably without incurring the costs of dedicated test calls.
Innovation Solution
A system comprising a transactions registry accessible by multiple operators and carriers, with a logging module that records calling and called numbers along with timing information, and a fraud detection and classification module that selects correlated entries to identify anomalies and classify calls into specific fraud types, utilizing blockchain technology for secure data sharing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If a wholesale carrier manages international traffic with multiple routing options, then cost efficiency is improved through dynamic routing choices, but transparency and visibility are lost making fraud detection difficult
Solution Approach 1:
The patent introduces a shared registry as an intermediary system that all carriers in the call chain contribute to and query. This mediator enables transparent visibility of the complete international call routing path without requiring changes to the cost-driven routing decisions themselves, allowing fraud detection while preserving cost efficiency
Solution Approach 2:
The system implements feedback by having each carrier log their handling of the international call to the shared registry. This creates a transparent audit trail that feeds back information about the actual call path to all participants, enabling detection of deviations from expected routing while maintaining the dynamic cost-efficient routing choices
2Measurement precision
If fraud detection systems use dedicated test calls to verify routing, then detection accuracy is improved, but operational costs and network load increase
Solution Approach 1:
The system uses self-service by leveraging the existing operational traffic and routing decisions of carriers themselves. Instead of introducing separate test calls, the fraud detection utilizes the actual call data that carriers are already generating and handling, eliminating the need for additional dedicated test calls while maintaining detection accuracy
Solution Approach 2:
The system performs preliminary action by having carriers log their call handling information in advance to the shared registry before fraud analysis is performed. This pre-positioning of data enables accurate fraud detection without requiring reactive test calls when fraud is suspected, reducing both costs and network load
3Loss of information
If carriers log detailed call information in a shared registry, then end-to-end visibility is improved for fraud detection, but system complexity and data management requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the fraud detection system into independent modular components: each carrier independently logs their segment of the call chain to the shared registry, and the fraud detection system independently queries and analyzes the relevant segments. This modular approach reduces system complexity while achieving end-to-end visibility through the combination of segmented data
Data Source
AI summary
A system for detecting fraud in international telecommunication traffic, includes: a transactions registry accessible by multiple operators and/or carriers; a logging module installed with multiple operators and/or carriers configured to add an entry in the transactions registry by an operator and/or carrier handling a call, wherein an entry comprises the calling number, the called number and timing information for the call; a fraud detection and classification module configured to: select from the transactions registry a group of correlated entries representing the call; compare the correlated entries to detect an anomaly due to fraud and to classify the call into a fraud type.