Multi-Level Backup Indexing for Rapid Data Restoration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing backup/restore applications face inefficiencies in searching and selecting the required backup files for data object restoration, particularly when dealing with large numbers of backup files created over time, leading to a significant reduction in the effectiveness of the restoration process.

Innovation Solution

Implementing a multi-level indexing system using a B-tree data structure to efficiently index and retrieve backup files, where full backup files are indexed in one level, differential files in a second level, incremental files in a third level, and transaction log files in a fourth level, allowing for rapid identification and restoration of data objects to specific times.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single-level index structure is used to store all backup files, then the system can maintain a simple indexing mechanism, but the time required to search and identify the necessary backup files for restoration increases significantly as the number of backup files grows

Engineering Contradiction:
Improveindexing structure complexityVSAvoidsearch time for backup files
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The patent divides the backup file indexing system into multiple hierarchical levels. The first level indexes full backup files, the second level indexes differential backup files associated with each full backup, the third level indexes incremental backup files, and the fourth level indexes transaction log backup files. This segmentation allows the system to quickly narrow down search scope by first identifying the relevant full backup, then drilling down through progressively more specific backup types, thereby reducing overall search time despite increased structural complexity.

Inventive Principle:
Principle #1Segmentation

2Productivity

If all backup files are stored and indexed at the same level, then the indexing structure remains simple, but the system efficiency deteriorates when restoring data objects to specific times because the system must search through all backup files regardless of type

Engineering Contradiction:
Improverestoration speedVSAvoidmulti-level indexing structure
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces a hierarchical dimension to the indexing structure, organizing backup files into four distinct levels based on their type and relationship. Instead of a flat single-level index, the system creates a tree-like structure where each level represents a different category of backup files. This dimensional organization enables the restoration process to systematically navigate from the most general level (full backups) to the most specific level (transaction logs), dramatically improving restoration speed by eliminating the need to search through all backup files of all types.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS11221983B1Multi-level indexing of backup files
Publication Date: 2022.01.11 EMC IP HLDG CO LLC
  • US11221983B1 patent drawing
  • US11221983B1 patent drawing
  • US11221983B1 patent drawing

AI summary

A system indexes a first backup file, created for a data object when a first time period expired, in a first level of a data structure. The system indexes a second backup file, created for the data object when a second time period expired, in a second level of the data structure, the second time period being shorter than the first time period. The system receives a request to restore the data object to a specified time. The system enables restoration of the data object by identifying a first indexed backup file associated with the specified time in the first level of the data structure, by identifying a part of the second level of the data structure corresponding to the first indexed backup file, and by identifying a second indexed backup file associated with the specified time in the part of the second level of the data structure.