File Level Restore Service for Virtual Machine Backups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud service providers lack file level restore (FLR) capabilities for Kernel-based Virtual Machine (KVM) or Xen hypervisor environments, requiring full image-level restoring of data volumes, which is inefficient and time-consuming.
Innovation Solution
A system and method for providing hypervisor-independent, storage-independent, and transport-independent FLR as a service, allowing users to browse and download individual files from a virtual machine backup without restoring the entire VM or disk, using a virtual disk file with an empty file system mounted as a root folder, and creating folders for each file system to enable selective file restoration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full image-level restoring is performed for data volumes in KVM and Xen based cloud environment, then complete data recovery is achieved, but restore time increases to hours and system resources are heavily consumed
Solution Approach 1:
The patent segments the virtual disk image into individual file system components, allowing selective mounting and restoration of specific file systems or folders rather than restoring the entire image. This enables users to restore only the necessary portions of data, significantly reducing restore time from hours to minutes or seconds while maintaining complete data recovery capability for the selected segments.
Solution Approach 2:
The patent extracts individual files or folders from the virtual disk image by mounting file systems directly from the backup image without full restoration. This extraction approach allows users to retrieve specific data elements while avoiding the time-consuming process of restoring the entire image, thus resolving the contradiction between complete recovery and fast restoration.
2Reliability
If full image-level restoring is performed for data volumes in KVM and Xen based cloud environment, then complete data recovery is achieved, but CPU, memory, disk I/O and network bandwidth consumption increase significantly
Solution Approach 1:
The patent divides the restoration process into segment-level operations where only selected file systems or folders are mounted and restored from the backup image. This segmentation prevents the heavy system resource consumption associated with full image-level restoration while ensuring complete recovery of the selected data segments, thus reducing CPU, memory, disk I/O and network bandwidth usage.
Solution Approach 2:
The patent extracts only the necessary files or folders from the backup image through direct file system mounting, avoiding the resource-intensive process of restoring the entire image. This extraction mechanism maintains data recovery completeness for the selected elements while significantly reducing system resource consumption during the restoration operation.
3Ease of operation
If FLR capability is implemented for KVM or Xen hypervisor, then selective file restoration is enabled, but system complexity increases due to hypervisor-specific SDK requirements
Solution Approach 1:
The patent implements a universal FLR solution that works across multiple hypervisors (KVM, Xen, VMware, Hyper-V) through a common approach of mounting virtual file systems from backup images. This universal method eliminates the need for separate hypervisor-specific SDK implementations, reducing system complexity while maintaining selective file restoration capability across different virtualization platforms.
Solution Approach 2:
The patent introduces an intermediary layer that handles the mounting and access of virtual file systems from backup images without requiring direct integration with hypervisor-specific SDKs. This intermediary approach simplifies the implementation by providing a unified interface for selective file restoration across different hypervisors, reducing the complexity associated with multiple platform-specific integrations.
4Reliability
If full disk restoration is performed, then all data is recovered, but user experience deteriorates due to long wait times and unnecessary data restoration
Solution Approach 1:
The patent segments the restoration process to allow users to select specific file systems or folders for restoration rather than forcing full disk restoration. This segmentation enables users to recover only the data they need, improving user experience by reducing wait times from hours to minutes or seconds while maintaining complete recovery capability for the selected segments.
Solution Approach 2:
The patent extracts and restores only the specific files or folders users need from the backup image through direct file system mounting. This extraction approach eliminates unnecessary data restoration, significantly improving user experience by reducing restore time and avoiding the restoration of irrelevant data while ensuring complete recovery of the selected elements.
Data Source
AI summary
A method and system for providing file level restore (FLR) service for restoring one or more files stored in a plurality of file systems in a backup of a first virtual machine is provided. The method creates a FLR session for a user, including: creating a virtual disk file in a second virtual machine providing the FLR service, the virtual disk file including an empty file system being mounted as a root folder of a virtual appliance in the second virtual machine; creating a respective folder for each of the plurality of file systems under the root folder; mounting each of the plurality of file systems to the respective folder; and mounting the root folder to a folder of the second virtual machine. The method restores the one or more files by the user through accessing the folder of the second virtual machine in the FLR session.