Workload Signature Grid Mapping for Storage Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current workload identification techniques in network storage systems are cumbersome and limited, as they rely on manual classification and generate complex, user-unfriendly reports, making it difficult for administrators to recognize shifts in workload types and maintain performance.
Innovation Solution
A graphical representation of workload classification is achieved by mapping workload signatures to a low-dimensionality grid using topographical mapping techniques like Kohonen self-organizing maps, allowing administrators to visualize and infer workload types and categories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual classification of workload types is used, then administrators can identify workload categories, but the process becomes cumbersome and difficult as workload complexity increases
Solution Approach 1:
The system automatically classifies workloads by analyzing workload signatures and mapping them to clusters without requiring manual administrator intervention. The workload manager continuously monitors and reclassifies workloads based on changing characteristics, making the system self-serve rather than relying on human operators.
Solution Approach 2:
The patent replaces the mechanical manual classification process with an automated computational system that uses workload signatures, clustering algorithms, and graphical representations to identify and categorize workloads, thereby eliminating the need for manual analysis of complex workload data.
2Loss of information
If detailed reports of workload types are generated, then comprehensive workload information is available, but the reports become complex and not user-friendly
Solution Approach 1:
The patent transforms detailed workload data from multiple dimensions into a two-dimensional graphical representation where workload signatures are mapped to clusters on a grid. This visual mapping preserves the essential characteristics of workloads while presenting them in an intuitive, easy-to-understand format that administrators can quickly interpret.
Solution Approach 2:
The system uses color-coded graphical representations to differentiate between various workload clusters and types. By assigning visual characteristics to different workload categories, the system enables administrators to quickly understand workload composition and changes without wading through complex textual reports.
3Adaptability or versatility
If traditional workload identification methods are used, then simple workload types can be recognized, but the system cannot effectively handle increasing workload complexity and parallel workload types
Solution Approach 1:
The patent creates a universal workload classification system that can handle diverse workload types including OLTP, DSS, and other parallel workloads through a single unified approach. The system uses workload signatures and clustering that adapt to any workload characteristics, making the identification system versatile across different workload scenarios without requiring separate methods for each type.
Data Source
AI summary
Technology is disclosed for graphically representing classification of workloads in a storage system. Workload classification is graphically represented to the user by mapping workload signatures of the workloads to a grid. When the workload signatures are mapped to the grid, a number of clusters are formed in the grid. Each of the clusters represents workloads of a particular category. Mapping the workload signature to the grid includes mapping a high-dimensionality workload signature vector to a low-dimensionality grid.