Tensor-Based Network Abnormality Detection in SDNs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting abnormalities in Software-Defined Networks (SDNs) rely on low-dimensional data and ignore geometric connection information, making them inadequate for meeting network security requirements.
Innovation Solution
The method involves acquiring reference and target tensors representing network traffic with dimensions of source, destination, and time, and determining a target core tensor based on a reference decomposition factor, with an abnormality detected if the difference between the target and reference core tensors exceeds a preset value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods use algorithms based on machine learning and statistical models to detect network abnormalities, then the detection process is simple to implement, but the detection accuracy is insufficient due to using only low-dimensional data and ignoring geometric connection information
Solution Approach 1:
The patent transforms network traffic data from conventional low-dimensional representation to high-dimensional tensor representation, incorporating geometric connection information of network devices. This dimensional expansion enables the detection method to capture complex network relationships and structural patterns that were previously invisible, thereby significantly improving detection accuracy while maintaining computational feasibility through tensor decomposition techniques.
2Reliability
If high-dimensional tensor data with geometric connection information is used for detection, then detection accuracy improves, but computational complexity and data processing requirements increase
Solution Approach 1:
The patent applies tensor decomposition to segment the high-dimensional network traffic tensor into multiple lower-order tensors, each representing different aspects of network behavior. This segmentation reduces the computational complexity of processing the complete high-dimensional data while preserving the essential geometric connection information and structural patterns needed for reliable anomaly detection.
Solution Approach 2:
The patent utilizes tensor decomposition to transform the high-dimensional tensor into a product of lower-dimensional tensors, effectively reducing the computational burden. This dimensional transformation maintains the critical geometric connection information while making the data more manageable for security detection algorithms, thus improving reliability without proportionally increasing system complexity.
Data Source
AI summary
Embodiments of the present disclosure provide a method, an apparatus, an electronic device, and a medium for detecting an abnormality in a network. The method for detecting an abnormality in a network includes acquiring a reference tensor and a target tensor representing traffic in the network, the reference tensor and the target tensor having at least dimensions of source, destination, and time of the traffic. The method further includes determining a target core tensor of the target tensor based on a reference decomposition factor of the reference tensor related to the dimensions of source and destination of the traffic. The method further includes determining that there is an abnormality in the network if a difference between the target core tensor of the target tensor and a reference core tensor of the reference tensor is greater than a preset value.