Workload Classification With GAF Images and LLMs

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The challenge in cloud computing environments is the inefficient detection of workload anomalies, which hinders timely resolution of technical issues, as manual inspection of workload logs is tedious and time-consuming.

Innovation Solution

The integration of workload gramian angular field (GAF) images and large language models (LLM) for automated workload classification, enabling efficient detection of anomalies by transforming workload logs into visual representations and leveraging advanced language understanding to classify workloads as anomalous or non-anomalous.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual inspection of workload logs is used, then detailed analysis can be performed, but the process is tedious and time-consuming

Engineering Contradiction:

Improveworkload log analysis efficiencyVSAvoidtime for manual inspection

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical inspection of workload logs with an automated system that uses Large Language Models (LLMs) to process and analyze logs. The LLM acts as an intelligent agent that automatically reads, understands, and classifies workload logs, eliminating the need for human manual inspection while maintaining detailed analysis capability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service analysis where the LLM autonomously performs workload log analysis without human intervention. The model automatically detects anomalies, classifies workloads, and provides insights, allowing the system to serve itself rather than requiring continuous human operational input.

Inventive Principle:

Principle #25Self-service

2Productivity

If automated classification systems are used, then analysis efficiency improves, but complexity of the system increases

Engineering Contradiction:

Improveworkload classification speedVSAvoidsystem architecture complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent employs a universal LLM-based system that can perform multiple functions: analyzing workload logs, detecting anomalies, classifying workloads, and providing explanations. This multi-functional approach consolidates what would otherwise require multiple separate specialized systems into a single versatile platform, managing complexity through functional integration rather than multiplication.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system manages complexity by dynamically adjusting LLM parameters such as temperature, max tokens, and prompt templates based on the specific analysis task. This parameter-based configuration allows the same underlying model to adapt to different workload scenarios without requiring complex architectural changes, simplifying system design while maintaining high performance.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250292542A1Workload classification using a large language model and gramian angular field images

Publication Date: 2025.09.18 DELL PROD LP

AI summary

A method for workload classification. The method includes: obtaining a workload log for a workload; producing, for the workload and based on the workload log, a workload gramian angular field (GAF) image set; and assigning, to the workload, a workload class at least based on the workload GAF image set. More specifically, embodiments described herein integrate visual representations of workload logs (in the form of GAF images) and advanced language understanding capabilities (offered by a large language model) to classify workload logs, and thus workloads, as anomalous or non-anomalous.