File Deletion Detection in Key-Value Databases for Virtual Backups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Protecting data in virtualized computing environments is inefficient due to the time and resources required for conventional backup and restoration methods, especially as data volume increases.
Innovation Solution
Implementing a deduplicated data repository and virtual environment manager to take snapshots of virtual machines, allowing for efficient incremental backups by tracking changed blocks and using replication to create independent full images, which can be used for rapid recovery and individual file retrieval without mounting the entire virtual disk image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional backup methods are used to protect data in virtualized environments, then data protection is achieved, but the process takes considerable time and resources
Solution Approach 1:
The patent segments the backup process by dividing virtual disk images into multiple chunks and processing them in parallel. Instead of backing up entire virtual disk images sequentially, the system divides the data into manageable segments that can be hashed, deduplicated, and transferred simultaneously, significantly reducing backup time while maintaining data protection integrity
Solution Approach 2:
The patent implements incremental backups that only transmit changed blocks rather than complete virtual disk images. By using change block tracking to identify only the portions of data that have changed since the last backup, the system performs partial actions that are sufficient for data protection while minimizing time and resource consumption
2Reliability
If conventional backup methods are used to protect data in virtualized environments, then data protection is achieved, but considerable resources are consumed
Solution Approach 1:
The patent uses cryptographic hashing to create digital fingerprints of data chunks instead of copying and transmitting actual data blocks. By hashing source data and comparing it with existing backup data, the system determines what needs to be transferred without moving redundant data, significantly reducing network bandwidth consumption and resource usage while ensuring data protection
Solution Approach 2:
The patent transforms the backup approach by changing from transmitting actual data blocks to transmitting only hash values for verification and minimal data blocks for restoration. This parameter change in the data representation and transmission method reduces resource consumption while maintaining the reliability of data protection through cryptographic verification
3Ease of operation
If virtual disk images are mounted for file retrieval, then individual files can be accessed, but the entire virtual disk image must be processed
Solution Approach 1:
The patent extracts individual file information from the virtual disk image by maintaining an index of file metadata (names, paths, hashes) separately from the actual data blocks. When file retrieval is needed, the system queries this index directly without mounting or processing the entire virtual disk image, extracting only the specific file data needed while leaving the rest of the image untouched
Solution Approach 2:
The patent introduces an intermediary indexing layer between the virtual disk image and the retrieval process. This index structure acts as a mediator that allows direct access to specific files through hash-based lookup, eliminating the need to mount or scan the entire virtual disk image and significantly reducing retrieval time while maintaining ease of file access
Data Source
AI summary
A method, article of manufacture, and apparatus for protecting data. In some embodiments, this includes receiving a changed block, reading a previous backup to get a file list in change block tracking, checking a file in a modified master file table list for a reference flag, wherein the modified master file table list is based on the changed block, and modifying a database entry based on the reference flag.