Cloud Resource RBAC Permission Discovery with Hierarchical LLM Prompts

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Evaluating effective permissions in complex role-based access control (RBAC) systems is challenging due to their hierarchical and granular nature, leading to errors and inconsistencies in manual or rule-based methods, which can result in security vulnerabilities and application outages.

Innovation Solution

A permissions analytics system utilizing a large language model (LLM) to analyze effective permissions by creating bespoke prompts that include context and user queries, allowing administrators to evaluate permissions through natural language responses.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual or rule-based methods are used to evaluate effective permissions, then the process is simple to implement, but errors and inconsistencies occur leading to security vulnerabilities

Engineering Contradiction:
Improveaccuracy of permissions analysisVSAvoidcomplexity of analysis system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces an LLM-based intermediary component that mediates between the RBAC system configuration and the permissions analysis output. This intermediary automatically interprets complex role definitions, action hierarchies, and permission assignments to generate accurate effective permissions evaluations, eliminating manual errors while maintaining system reliability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual mechanical evaluation methods with an automated LLM-based system. The LLM processes role definitions, action definitions, and permission assignments through natural language understanding, substituting human manual analysis with intelligent automated processing that maintains high accuracy while reducing operational complexity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If LLM-based analysis is implemented, then reliability and accuracy of permissions analysis improve, but computational resources and system complexity increase

Engineering Contradiction:
Improveprecision of effective permissions evaluationVSAvoidcomputational resource consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the permissions analysis process into distinct components: role definition processing, action definition processing, permission assignment evaluation, and effective permissions generation. The LLM processes each segment separately based on the specific analysis needs, reducing overall computational resource consumption while maintaining high precision in the final evaluation results.

Inventive Principle:
Principle #1Segmentation

3Reliability

If detailed action definitions are provided to LLM, then analysis accuracy improves, but prompt size and processing time increase

Engineering Contradiction:
Improveconsistency of permissions evaluationVSAvoidprocessing time for permissions analysis
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary processing of action definitions by organizing them into hierarchical structures and pre-processing role configurations before LLM analysis. This preliminary action prepares the data in an optimized format that the LLM can process efficiently, maintaining consistent and accurate permissions evaluation while reducing the actual processing time during execution.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4625228A1Automatic effective permissions discovery for cloud resources
Publication Date: 2025.10.01 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4625228A1 patent drawingFigure 1
  • EP4625228A1 patent drawingFigure 2
  • EP4625228A1 patent drawingFigure 3

AI summary

Examples analyze effective permissions in a role-based access control system. A prompt is created for a large language model (LLM). The prompt includes a role definition for a role of a role-based access control system, and action definitions. The role definition for the role includes an action and effective permissions text describing a summary of permission limitations provided by the role. The action definitions are provided in a hierarchical format. Query text is added to the prompt. The query text includes a question about effective permissions associated with the role. The prompt is submitted to the LLM, thereby generating response text from the LLM. The response text is displayed to a user.