ML Classification of Object Store Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating warnings and recommendations for object store workloads rely on explicitly-coded heuristics, which are time-consuming, subjective, and not individualized, making them inefficient for modern large-scale object stores.
Innovation Solution
A system that uses machine learning classification based on resource utilization descriptors to provide tailored warnings and recommendations, leveraging a machine learning model to analyze CPU, memory, and network bandwidth utilization levels, enabling quick, objective, and scalable insights.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If explicitly-coded heuristics are used to generate warnings and recommendations, then the system can provide structured analysis, but the process becomes time-consuming and not individualized
Solution Approach 1:
The patent replaces the mechanical system of explicitly-coded heuristics with a machine learning model that automatically learns patterns from historical workload data. The ML model processes resource utilization descriptors and generates classification labels without manual rule configuration, thereby reducing time consumption while maintaining analysis accuracy through data-driven insights.
Solution Approach 2:
The system enables self-service by allowing the machine learning model to autonomously generate individualized warnings and recommendations based on learned patterns from historical data. The model automatically adapts to different workload characteristics without requiring manual heuristic configuration for each case, providing both speed and personalization.
2Stability of the object's composition
If explicitly-coded heuristics are used, then consistent rules can be applied, but the approach is subjective and not individualized to specific workloads
Solution Approach 1:
The patent changes the fundamental parameter of analysis from fixed heuristic rules to dynamic machine learning classification labels. The ML model learns optimal parameters from historical workload data and applies them individually to each workload, providing both consistency through standardized ML processing and adaptability through data-driven parameter optimization for each specific workload type.
Solution Approach 2:
The system segments workloads into distinct classification categories based on their resource utilization patterns. The machine learning model divides the continuous workload space into discrete classes (e.g., different workload types or performance levels), enabling individualized analysis for each segment while maintaining overall system consistency through uniform classification methodology.
3Ease of manufacture
If traditional heuristic methods are used for workload analysis, then implementation is straightforward, but the system cannot scale efficiently to large object stores
Solution Approach 1:
The patent implements a universal machine learning model that can handle diverse workload types across the entire object store infrastructure. The same ML model architecture processes different workload characteristics (CPU, memory, network, storage utilization) and generates appropriate classification labels, enabling scalable deployment across large numbers of workloads while maintaining ease of implementation through a single standardized system.
Data Source
AI summary
Systems/techniques that facilitate machine learning classification of object store workloads are provided. In various embodiments, a system can access a resource utilization descriptor associated with an object store. In various aspects, the system can generate, via execution of a machine learning model (e.g., a deep learning neural network, a random forest model), a classification label based on the resource utilization descriptor. In various instances, the system can perform one or more electronic actions based on the classification label. In various cases, the classification label can indicate/identify a computing fault of the object store, can indicate/identify resources of the object store that are being underutilized and/or overutilized, and/or can indicate whether a workload corresponding to the resource utilization descriptor could be properly transplanted to a different object store. Accordingly, the one or more electronic actions can include generating warnings/recommendations regarding such computing fault, such underutilized/overutilized resources, and/or such transplantation.