Virtual Hard Disk Backup Prefetching for Storage Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Backup storages face bottlenecks due to high volumes of activity, particularly from applications performing random reads, which can lead to delayed data retrieval and unsatisfactory performance in meeting time-sensitive IO demands, as they are not application-aware and are optimized for writes rather than efficient restorations.
Innovation Solution
The implementation of a method that uses VHDX format to enable native mounting and reading/writing of backup files without kernel drivers or virtual file systems, allowing for item-level restorations and prefetching of backup data, thereby reducing the load on both applications and backup storage by caching data for faster access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If backup storage performs random reads for application data access, then data retrieval can be attempted, but the backup storage becomes bottlenecked and command timeouts increase due to high activity volumes and write-optimization
Solution Approach 1:
The patent segments the backup data access process by creating a virtual file system that separates metadata operations from actual data retrieval. The virtual file system handles random read requests locally using cached metadata, while only transferring actual data blocks when necessary. This segmentation reduces the burden on backup storage by eliminating unnecessary random read operations.
Solution Approach 2:
The system performs preliminary actions by pre-caching backup metadata and data blocks in the virtual file system before they are actually needed. The metadata is loaded in advance, allowing the system to predict and prepare for upcoming data access patterns. This preliminary caching action prevents bottlenecking during actual data retrieval operations.
2Ease of operation
If entire backup data is copied to host for restoration, then complete data access is achieved, but the time and resources required increase significantly
Solution Approach 1:
The patent extracts only the necessary portions of backup data rather than copying the entire backup set to the host. The virtual file system identifies and retrieves only the specific data blocks required for the application's access needs, leaving the remainder of the backup data on the backup storage. This extraction approach dramatically reduces both time and resource requirements.
Solution Approach 2:
The system performs partial action by retrieving and caching only the subset of backup data that is actually needed for current and predicted future access operations. Rather than performing the excessive action of copying all backup data, the system dynamically determines the minimum necessary data transfer required to satisfy application demands.
3Productivity
If backup storage is optimized for writes, then backup efficiency is improved, but read operations and restorations become slow and bottlenecked
Solution Approach 1:
The patent introduces a virtual file system as an intermediary layer between the backup storage and the application. This intermediary maintains cached metadata and data blocks locally, translating application read requests into efficient operations that do not require direct access to the write-optimized backup storage. The intermediary handles the mismatch between write-optimized storage characteristics and read-oriented application demands.
Data Source
AI summary
Techniques described herein relate to a method for performing data protection of file system data on a host. The method includes identifying, by a data protection agent, a backup access event associated with a backup of a file system stored on a backup storage; in response to identifying the backup access event: obtaining backup metadata associated with the backup from a data protection manager; generating a placeholder file system using the backup metadata and storing the placeholder file system in virtual hard disk file; loading the virtual hard disk file on a target application; performing, after the loading, prefetching of backup data using the virtual hard disk file and the backup metadata to store the backup data in a cache; and performing backup access services using the virtual hard disk file, the backup metadata, and the cache.