Cloud Access Governance Using Dynamic Clustering for Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud-based access privilege governance systems lack efficient methods to dynamically determine the number of clusters for peer group analysis, leading to performance issues and inaccurate anomaly detection in large-scale identity and access data sets.
Innovation Solution
The system retrieves identity access data, encodes it as binary vectors, uses min-wise independent permutations locality sensitive hashing to determine a distinct identity access count, normalizes the cluster count, and performs peer group analysis using k-means clustering to identify access profile anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional clustering methods are used for peer group analysis in cloud-based access privilege governance, then the system can process identity and access data, but the performance deteriorates and accuracy decreases when handling large-scale data sets
Solution Approach 1:
The patent segments the large-scale identity and access data into smaller manageable clusters using k-means clustering algorithm. By dividing the vast data set into distinct groups based on similarity metrics, the system can process each cluster efficiently rather than handling the entire data set as one monolithic structure, thus improving productivity while managing large data volumes.
Solution Approach 2:
The patent transforms the high-dimensional identity and access data into a lower-dimensional representation by determining a reduced number of clusters. This dimensionality reduction allows the system to maintain the essential characteristics of the data while significantly reducing computational complexity, enabling efficient processing of large-scale data sets.
2Measurement precision
If the number of clusters is not dynamically determined, then the system structure remains simple, but the accuracy of anomaly detection deteriorates
Solution Approach 1:
The patent implements dynamic determination of the number of clusters by using the identity access count as a variable parameter. Instead of using a fixed number of clusters, the system adapts the cluster count based on the actual data characteristics and access patterns, thereby improving anomaly detection accuracy while managing system complexity through automated adaptation.
Solution Approach 2:
The system performs self-adjustment by automatically determining the optimal number of clusters based on the identity and access data provided. The normalized identity access count serves as a self-derived parameter that guides the clustering process, eliminating the need for manual configuration and external intervention to achieve accurate anomaly detection.
3Loss of time
If manual access reviews are conducted without automated clustering, then the system requires minimal computational resources, but the time consumption and labor requirements increase significantly
Solution Approach 1:
The patent replaces manual mechanical review processes with automated computational clustering algorithms. By substituting human-driven access reviews with k-means clustering and peer group analysis, the system significantly reduces review time while the computational power required is manageable through efficient algorithm implementation and optimized cluster determination.
Data Source
AI summary
Embodiments provide cloud based access privilege governance. Embodiments retrieve identity access data and encode the identity access data as a plurality of binary vectors. Embodiments determine a distinct identity access count as a cluster count, normalize the cluster count, and perform peer group analysis using the normalized cluster count.