Tape Backup Data Organization for Reduced Seek Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for backing up a file system with file-level deduplication to a tape device result in frequent seek operations, leading to elongated restore times due to not considering the characteristics of the tape device and the deduplicated state being released during backup.
Innovation Solution
A system where shared data associated with deduplicated data is stored on the tape device before deduplicated data, allowing for sequential reading and reducing seek operations by prioritizing the storage and retrieval of shared data, thereby maintaining the deduplicated state during backup and restore processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional backup methods are used on tape devices for deduplicated data, then backup can be performed, but frequent seek operations occur leading to elongated restore time
Solution Approach 1:
The patent applies preliminary action by storing shared data (which is referenced by multiple deduplicated files) at the beginning of the backup data on the tape device before storing the deduplicated file data. This preliminary arrangement ensures that during restoration, the shared data is already in position and can be immediately utilized, eliminating the need for frequent seek operations and reducing restore time.
2Quantity of substance
If deduplicated data is backed up to tape device, then storage efficiency is improved, but the deduplicated state is released during backup
Solution Approach 1:
The patent uses copying by creating a specialized backup format that copies the reference structure of deduplicated data to the tape device. Instead of copying actual file data multiple times, it copies metadata indicating which files reference which shared data blocks, thereby maintaining the deduplicated state in the backup and reducing the amount of data actually stored on tape.
3Loss of time
If shared data is stored before deduplicated data on tape, then seek operations are minimized, but data organization becomes more complex
Solution Approach 1:
The patent applies segmentation by dividing the backup data into distinct segments: a first segment containing shared data and a second segment containing deduplicated file data. This segmentation creates a clear organizational structure that simplifies the restore process by allowing the system to sequentially access shared data first, then access specific file portions without random seeking, thereby reducing seek operations despite the organized structure.
Data Source
AI summary
It is provided a computer system, comprising a server and first and second storage systems. The first storage system stores deduplicated data sharing at least a part of data with other data, shared data shared by a plurality of pieces of the deduplicated data, and first type data representing a type of the stored data including the deduplicated data and the shared data. The deduplicated data is associated with the shared data by a pointer to the shared data, and includes differential data indicating a difference from the shared data. The server creates second type data representing a type of the data stored in the second storage system from the first type data. The second storage system stores the shared data associated with the deduplicate data at a reading position before a position at which the deduplicated data is read in sequential reading and stores the second type data.