Historical Index Metadata Query for Snapshot Volumes

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing methods for searching metadata across multiple snapshot volumes and base volumes in cloud storage are inefficient, leading to large time overheads due to the need for level-by-level searching and high metadata duplication, which increases storage space requirements.

Innovation Solution

A metadata query method that utilizes historical index information to quickly identify the location of metadata by correlating data block identifiers with their creation time sequences, allowing direct access to target volumes and reducing the need for level-by-level searches, thereby improving search efficiency and optimizing storage space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If level-by-level searching is performed across multiple snapshot volumes and base volumes, then metadata can be found, but time overheads for searching are relatively large

Engineering Contradiction:
Improvemetadata retrieval accuracyVSAvoidsearching time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent pre-calculates and stores time sequence identifiers and volume identifier mappings in historical index information before queries occur. This preliminary action allows the system to directly retrieve metadata location information without performing time-consuming level-by-level searches at query time, thus resolving the contradiction between search accuracy and searching time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces historical index information as an intermediary data structure that stores pre-computed mappings between data block identifiers, time sequence identifiers, and volume identifiers. This intermediary layer acts as a cache or lookup table that bridges the query interface and the actual metadata storage, eliminating the need for sequential searching through multiple volumes.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If multiple copies of metadata are maintained for snapshot volumes, then metadata availability is improved, but storage space requirements increase

Engineering Contradiction:
Improvemetadata availabilityVSAvoidmetadata storage space
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent merges the functionality of storing multiple metadata copies with a centralized historical index structure. Instead of duplicating full metadata across multiple volumes, the system combines volume identifier mappings and time sequence information into a single historical index, reducing total storage requirements while maintaining the ability to retrieve metadata from appropriate volumes.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a lightweight copy or reference structure (historical index) that contains only essential mapping information rather than full metadata copies. This selective copying approach maintains metadata availability by storing volume location information without duplicating the actual metadata content, thus reducing storage space requirements.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3678015B1Metadata query method and device
Publication Date: 2022.11.09 HUAWEI TECH CO LTD
  • EP3678015B1 patent drawingFigure 1~2
  • EP3678015B1 patent drawingFigure 3~5
  • EP3678015B1 patent drawingFigure 6~7

AI summary

In the field of data storage, a metadata query method and an apparatus are disclosed, to improve metadata searching efficiency. The method is applied to a linked snapshot, and includes: receiving a metadata query request, where the metadata query request includes a volume identifier of a first snapshot volume and a data block identifier (301); obtaining a first time sequence identifier from the first snapshot volume based on the volume identifier of the first snapshot volume, where the first time sequence identifier is used to indicate a creation time sequence of the first snapshot volume (302); querying historical index information based on the data block identifier and the first time sequence identifier, where the historical index information includes a correspondence between a query data block identifier and historical query snapshot information (303); and when the data block identifier exists, and the first time sequence identifier falls within a corresponding query time sequence interval, obtaining a corresponding target volume identifier from the historical index information, and obtaining, from a second snapshot volume indicated by the target volume identifier, address metadata corresponding to the data block identifier (304).