Snapshot Driver Bitmap Tracking for Copy-on-Write Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing snapshot technologies incur unnecessary copy-on-write overhead even for blocks that have already been read by consumer applications, as they continue to copy data for potential future writes, leading to inefficiencies in backup and replication processes.
Innovation Solution
Implementing a bitmap system where the snapshot driver tracks which blocks are of interest to the consumer application, allowing the driver to avoid copy-on-write operations for blocks that have been read or are not part of the snapshot, thereby reducing the number of copy-on-writes and associated overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If copy-on-write operations are performed for all blocks in the volume during snapshot lifetime, then data consistency is maintained, but unnecessary overhead is incurred for blocks already read by consumer applications
Solution Approach 1:
The patent applies local quality by treating different blocks differently based on their specific characteristics. Blocks that have been read by consumer applications are identified and excluded from copy-on-write operations, while blocks that haven't been read continue to undergo copy-on-write. This selective approach maintains data consistency for necessary blocks while avoiding unnecessary overhead for already-read blocks.
Solution Approach 2:
The patent implements preliminary action by pre-identifying blocks that have been read by consumer applications before copy-on-write operations are triggered. By maintaining tracking information about which blocks have been accessed, the system can proactively determine which blocks should be excluded from future copy-on-write operations, preventing unnecessary overhead before it occurs.
2Duration of action of moving object
If snapshots are maintained for long durations to support backup/replication processing, then data can be persisted, but copy-on-write overhead accumulates over time
Solution Approach 1:
The system performs preliminary identification of blocks that have been read during the snapshot lifetime, maintaining tracking information that accumulates over time. This allows the system to progressively identify more blocks eligible for exclusion from copy-on-write operations as the snapshot duration extends, reducing accumulated overhead while preserving the ability to maintain snapshots for extended periods.
3Reliability
If copy-on-write operations are performed for every write attempt to a block, then snapshot integrity is preserved, but processing time is wasted on unnecessary copies
Solution Approach 1:
The patent applies local quality by implementing block-level differentiation based on read status. Each block is evaluated individually to determine whether it requires copy-on-write protection, rather than applying a uniform policy to all blocks. This selective approach preserves snapshot integrity for blocks that need it while eliminating unnecessary processing time for blocks that have already been read.
Solution Approach 2:
The system performs preliminary checks using tracking information to determine whether a block requires copy-on-write operations before actually performing the copy. By pre-identifying blocks that have been read by consumer applications, the system可以避免 unnecessary copy operations and their associated processing time delays.
Data Source
AI summary
An attempt to write to a block of data in a main volume of data is detected. An indicator associated with the block of data is accessed before a copy-on-write operation to a snapshot volume is performed for the block of data. The indicator is used to determine whether the copy-on-write operation is to be performed for the block of data.