Block-Based Backup for Volumes Exceeding 64 TB
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current block-based backup systems are limited to supporting volumes up to 64 TB, forcing them to revert to file-level backups for larger volumes, which is inefficient and limits their availability.
Innovation Solution
The proposed solution involves partitioning each volume into 64 TB sub-volumes, storing each sub-volume as an individual VHDx, and using a volume partition process that knows the block number/offset to store blocks in the corresponding VHDx, allowing seamless backup of volumes of any size using a new volume-partitioned block-level backup framework.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If block-based backup is used for volumes greater than 64 TB, then backup efficiency and speed are improved, but the system cannot support volumes larger than 64 TB due to VHDx container limitations
Solution Approach 1:
The patent divides a large volume greater than 64 TB into multiple smaller volumes, each within the 64 TB limit. It creates multiple VHDx containers (first VHDx container, second VHDx container, etc.) to store backup images of these divided volumes. This segmentation allows the system to handle volumes of any size by breaking them down into manageable chunks that fit within container limitations.
Solution Approach 2:
The patent implements a hierarchical structure where multiple VHDx containers are nested within a parent backup image structure. Each VHDx container contains backup data for a specific volume segment, and all containers are managed together as part of a unified backup system. This nesting allows the system to maintain block-based backup efficiency while supporting arbitrary volume sizes.
2Adaptability or versatility
If file-level backup is used for volumes greater than 64 TB, then the system can handle any volume size, but backup performance deteriorates due to file system overhead
Solution Approach 1:
Instead of using file-level backup which processes files individually through the file system, the patent segments the volume at the block level and creates block-based backup images in VHDx containers. This segmentation occurs at the volume level rather than file level, allowing the system to bypass file system overhead while still handling volumes of any size through multiple segmented containers.
3Device complexity
If a single VHDx container is used for backup, then the backup structure is simple, but the maximum volume size is limited to 64 TB
Solution Approach 1:
The patent maintains relative simplicity by using standard VHDx container formats for each segment, but extends capability by creating multiple such containers. Each container remains individually simple and compatible with existing tools, while the collection of containers collectively supports unlimited volume sizes. The system manages multiple simple structures rather than requiring a single complex structure.
Solution Approach 2:
The patent makes the backup system universal by designing it to handle both small volumes (single VHDx container) and large volumes (multiple VHDx containers) using the same block-based backup mechanism. The system can adapt the number of containers based on volume size, providing a unified approach that works for any volume size without requiring different backup methods.
Data Source
AI summary
Embodiments are directed to a method of performing block based backups for large-scale source volumes, by obtaining a total source volume size to determine if the source volume exceeds 64 TB, creating a number of sub-volumes to divide the total source volume into discrete sub-volumes of 64 TB each, grouping the discrete sub-volumes into a container, and storing, upon a read operation for a block based backup operation, each block of the source volume into a corresponding sub-volume of the sub-volumes. The method determines a block number and offset of each block and derives the corresponding sub-volume from the offset. Each sub-volume represents an individual saveset of the backup operation, and the container manages the individual saveset for the block based backup operation.