Machine Learning Confidence for Storage Volume Reclamation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In large enterprise or cloud-based storage environments, identifying reclaimable storage volumes is a difficult, manual, and time-consuming process due to imperfect communication between storage administrators and application teams, often resulting in unused volumes consuming resources, with existing methods like reporting orphan volumes or querying inactive volumes leading to false positives and inefficiencies.

Innovation Solution

A method using machine learning techniques and environmental metadata to identify reclaimable storage volumes by training a system with a set of storage parameters and user decisions, providing a graphical user interface to present candidate volumes based on a calculated confidence level, allowing for continuous refinement and adjustment based on user context and permissions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional reporting methods for orphan volumes are used, then the process is simple to implement, but the accuracy of identifying reclaimable volumes is low due to errors and omissions

Engineering Contradiction:

Improveaccuracy of identifying reclaimable volumesVSAvoidcomplexity of identification system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary machine learning model that acts as a mediator between traditional identification methods and reclamation decisions. This model processes multiple input features (host mappings, I/O activity, volume attributes) and produces a confidence score, thereby improving accuracy without requiring direct complex correlation logic in the reclamation system itself.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual mechanical reporting and correlation processes with an automated machine learning system. Instead of storage administrators manually reviewing reports or using complex correlation tools, the system automatically analyzes volume characteristics and predicts reclamation probability, substituting human effort and mechanical report generation with intelligent automation.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If manual review of storage volumes is performed, then false positives can be avoided, but the time and resources required are excessive

Engineering Contradiction:

Improvereliability of reclamation decisionsVSAvoidtime for reclamation operations

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent applies partial action by not requiring complete manual review of all volumes. Instead, the machine learning model processes all candidate volumes automatically and presents only those with high confidence scores for potential reclamation. This partial automation approach maintains reliability by filtering out false positives while reducing time loss by eliminating manual review of obvious candidates.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system performs self-service by automatically identifying and prioritizing reclaimable volumes without requiring storage administrators to manually examine each volume. The machine learning model independently evaluates volumes, generates confidence scores, and presents recommendations, allowing the system to serve itself in the initial identification phase while human experts focus on final verification of high-confidence candidates.

Inventive Principle:

Principle #25Self-service

3Productivity

If querying for inactive volumes is used, then potential reclamation candidates can be found, but false positives increase due to volumes active on monthly or quarterly basis

Engineering Contradiction:

Improveefficiency of reclamation identificationVSAvoidaccuracy of reclamation candidates

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent changes parameters by moving beyond simple I/O activity thresholds to a multi-parameter analysis including host mappings, volume attributes, access patterns, and temporal characteristics. The machine learning model evaluates multiple parameters simultaneously and dynamically adjusts their weightings, allowing it to distinguish between truly inactive volumes and those with periodic batch processing activity, thereby improving precision while maintaining productivity.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10445659B2Machine learning for determining confidence for reclamation of storage volumes

Publication Date: 2019.10.15 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

A method, system and computer product for performing storage maintenance is described. A training set for storage volume reclamation is received. The training set includes a set of storage parameters, each set of storage parameters corresponds to a respective candidate storage volume of a set of candidate storage volumes. The training set also includes a set of user decisions made whether a respective candidate storage volume is reclaimable. The training set is used to train a machine learning system to recognize common features of reclaimable candidate storage volumes. A set of candidate storage volumes is provided for potential reclamation, each with a set of storage parameters. A graphical user interface presents respective members of the set of candidate storage volumes for reclamation if a confidence level is calculated that the respective candidate storage volume is reclaimable exceeds a threshold.