Backup Virtualization via Application and Volume Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing backup and recovery systems face challenges in providing rapid, simple, and accurate data recovery while minimizing overhead on applications, often resulting in inconsistent backups due to ongoing system modifications during the backup process.
Innovation Solution
The method involves generating and storing backup objects that represent storage volumes at a specific point in time, using application, object, and volume maps to identify and store data locations, allowing for consistent block-level backups that reduce system overhead and facilitate quick recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If application-assisted backup operations are used to back up logical objects, then backup completeness is improved, but system overhead increases and consistency deteriorates
Solution Approach 1:
The patent introduces a backup system that acts as an intermediary between the application and storage volumes. The backup system directly accesses storage volumes to obtain backup data without requiring application assistance, thereby reducing system overhead while maintaining backup completeness. The storage volumes themselves serve as intermediaries that provide data location information to the backup system.
Solution Approach 2:
The storage volumes are designed to be self-describing, containing metadata that automatically identifies data locations and relationships. This self-service capability allows the backup system to navigate and copy data without requiring application-level assistance, reducing the burden on applications while ensuring complete backup coverage.
2Reliability
If application-assisted backup operations are used to back up logical objects, then backup completeness is improved, but backup consistency deteriorates
Solution Approach 1:
The backup system performs preliminary actions by first identifying all data locations and relationships in the storage volumes before initiating data copying. The system uses metadata from storage volumes to map out the complete data structure in advance, ensuring that all related files and objects are captured together, thus maintaining backup consistency.
Solution Approach 2:
The backup system acts as an intermediary that directly interfaces with storage volumes to obtain both data and its location metadata. This direct access allows the backup system to capture the complete state of storage volumes at a point in time without being subject to application modifications during the backup process, ensuring consistency.
3Quantity of substance
If block-level backup images are stored without virtualization, then storage efficiency is improved, but recovery flexibility and speed deteriorate
Solution Approach 1:
The patent segments backup data into block-level images that are individually stored and managed. Each block-level image represents a specific portion of the storage volume at a point in time. This segmentation allows the backup system to store data efficiently while enabling rapid recovery by selecting and restoring only the specific blocks needed, rather than restoring entire volumes.
Solution Approach 2:
The patent introduces a virtualization dimension by creating a backup repository that virtualizes block-level images. This virtualization layer adds an organizational dimension that maps logical data structures to physical block storage, enabling both efficient storage utilization and fast recovery operations through intelligent data location and retrieval.
4Quantity of substance
If block-level backup images are stored without virtualization, then storage efficiency is improved, but application-specific recovery capability deteriorates
Solution Approach 1:
The patent introduces a virtualization dimension that creates multiple levels of abstraction over block-level images. The backup repository virtualizes data by creating application-specific views that map to underlying block-level images. This allows the system to maintain efficient block-level storage while providing application-specific recovery capabilities through the virtualization layer.
Solution Approach 2:
The backup repository is designed with multi-functionality, serving both as an efficient storage mechanism for block-level images and as a flexible recovery system that can restore data at multiple levels (block-level, file-level, application-level). The virtualization layer enables the same infrastructure to support diverse recovery scenarios without sacrificing storage efficiency.
Data Source
AI summary
A system uses one or more processors and an interface to a backup repository to generate a set of backup objects for a point in time, store the set of backup objects in the backup repository, and store mapping information usable to retrieve appropriate backup objects when desired. A first map relates one or more applications to objects used by the applications. A second map relates the objects to corresponding data blocks or other data locations on volumes. A third map relates the data locations to corresponding backup objects for the point in time.