Collaborative Filtering for IHS Issue Resolution Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Administering and diagnosing technical issues in large Information Handling Systems (IHS) within data centers is challenging due to the complexity of configuring and understanding the interrelated components, requiring administrators to spend significant time analyzing telemetry and log data to identify and resolve issues.
Innovation Solution
Implementing a method that uses collaborative filtering and linear regression to generate a ranked list of recommended resolutions for reported issues by identifying similar issues within a repository, utilizing vectorized representations and administrator feedback to prioritize solutions based on time proximity, cost, and effectiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If administrators manually analyze telemetry and log data to diagnose issues, then diagnostic accuracy can be achieved, but the time and effort required increases significantly
Solution Approach 1:
The patent introduces an intermediary system that includes a repository of stored issues and resolutions, and an issue analysis module that automatically matches reported issues with similar historical issues. This intermediary system bridges the gap between raw telemetry data and administrator decision-making, providing pre-analyzed resolution recommendations that maintain diagnostic accuracy while dramatically reducing the time administrators need to spend on manual analysis.
Solution Approach 2:
The system creates copies of historical issue patterns and resolutions, storing them in a repository that can be quickly queried and matched against new issues. By copying and reusing proven resolution approaches from similar historical cases, the system eliminates the need for administrators to re-analyze every issue from scratch, thereby reducing diagnostic time while maintaining effectiveness through proven solutions.
2Measurement precision
If administrators manually evaluate IHS configuration and interface elements to diagnose issues, then accurate problem identification can be achieved, but the complexity of the task increases
Solution Approach 1:
The system extracts the complex task of configuration analysis and pattern recognition from the administrator's workflow and transfers it to an automated issue analysis module. This module automatically evaluates IHS configuration, telemetry data, and log files, extracting only the essential diagnostic information and resolution recommendations. This extraction reduces the perceived complexity for administrators while maintaining accurate problem identification through systematic automated analysis.
3Measurement precision
If comprehensive telemetry and log data are collected from all IHSs, then better issue diagnosis can be achieved, but the amount of data to be processed increases
Solution Approach 1:
The system segments the large volume of telemetry and log data into structured, queryable units stored in a repository. Each historical issue is segmented with its associated telemetry patterns, log characteristics, and resolution outcomes. This segmentation allows the issue analysis module to efficiently query and compare only relevant segments when diagnosing new issues, rather than processing all raw data, thereby maintaining comprehensive analysis quality while reducing processing complexity through organized data structures.
Data Source
AI summary
Systems and methods are provided for providing recommended resolutions to issues occurring during operation of an Information Handling System (IHS). Similar issues, each associated with resolution information specifying one or more recommended resolutions, are identified within an issue repository. Collaborative filtering is utilized on the similar issues and associated resolution information to generate a first ranked list. Collaborative filtering is also utilized on a subset of the similar issues and associated resolution information to generate a second ranked list of the subset of similar issues and associated resolution information, where the subset of similar issues are each selected based on being within a time proximity to the reported issue. A linear regression is utilized on the first ranked list and the second ranked list to generate a ranked list of recommended resolutions to the reported issue that is then provided to an administrator.