Metadata-Guided Data Prestaging for Thin-Volume Read Latency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional block storage systems experience increased data latency due to the disconnect between logical and physical addresses, particularly in thin volumes where data is not sequentially allocated, leading to inefficient data retrieval.

Innovation Solution

A computer-implemented method that analyzes metadata to determine the sequentiality of physical addresses and selectively prestages data from those addresses, refraining from prestaging when it would harm performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If data is stored in thin volumes with logical-to-physical address mapping, then storage flexibility and space optimization are improved, but data latency increases due to non-sequential physical allocation

Engineering Contradiction:

Improvestorage flexibilityVSAvoiddata latency

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of metadata to identify sequential physical address patterns before data retrieval operations. By pre-determining which data blocks are sequentially allocated, the system can optimize read operations in advance, reducing latency without sacrificing storage flexibility

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts data retrieval strategies based on the sequentiality of physical addresses. When sequential patterns are detected, optimized read paths are applied; when non-sequential, standard mapping procedures are used. This dynamic adaptation resolves the contradiction by making the system responsive to actual data distribution patterns

Inventive Principle:

Principle #15Dynamics

2Productivity

If cache prefetch algorithms are applied to all read requests, then read efficiency may improve for sequential data, but performance degrades when data is non-sequential due to wasted prefetch operations

Engineering Contradiction:

Improveread efficiencyVSAvoidwasted prefetch operations

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The system applies different prefetching strategies to different data regions based on their sequentiality characteristics. Sequential data regions receive aggressive prefetching treatment, while non-sequential regions use conservative or no prefetching. This localized quality adjustment optimizes read efficiency for sequential data while avoiding wasted operations on non-sequential data

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system uses metadata analysis to provide feedback about physical address sequentiality to the prefetch algorithm. This feedback mechanism enables the system to adapt prefetch behavior based on actual data patterns, improving read efficiency when appropriate and avoiding wasted operations when data is non-sequential

Inventive Principle:

Principle #23Feedback

3Productivity

If metadata analysis is performed for every read request to determine sequentiality, then selective prestaging can optimize performance, but processing overhead increases

Engineering Contradiction:

Improveselective prestaging performanceVSAvoidprocessing overhead

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system performs partial metadata analysis by focusing only on the specific portions of metadata needed to determine sequentiality for the requested data blocks. Rather than analyzing complete metadata structures, the system extracts only the necessary address mapping information, reducing processing overhead while maintaining the ability to perform selective prestaging

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12566708B2Using metadata to selectively perform data prestaging

Publication Date: 2026.03.03 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

A computer-implemented method, according to one approach, includes: receiving a read request for data in memory and analyzing metadata from volumes that contain the requested data. The metadata maps sequential logical addresses of the requested data to their corresponding physical addresses. The computer-implemented method also includes determining whether the corresponding physical addresses storing the requested data are sequential. In response to determining that at least a predetermined number of the corresponding physical addresses are sequential, the requested data is prestaged from the sequential ones of the physical addresses. However, in response to determining that fewer than the predetermined number of the corresponding physical addresses are sequential, the computer-implemented method includes intentionally refraining from prestaging any of the requested data from the physical addresses.