Lazy Data Compression for Block-Level Storage Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data compression methods in data storage devices face inefficiencies due to resource-intensive decompress and recompress operations, especially when dealing with large compression domains, leading to performance bottlenecks and requiring higher-level entity management, which is not feasible in all storage systems.
Innovation Solution
Implementing lazy compression by designating data storage areas as compressed logical units where data is written in uncompressed form and compressed only upon detection of specific trigger conditions, allowing for block-based compression domains and deferring recompression until necessary, thereby optimizing system performance and resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If compression algorithms search the entirety of the data to be compressed to find the largest repeating sequences, then compression efficiency improves, but system performance deteriorates because the entire compression domain must be decompressed, modified, and recompressed when any portion changes
Solution Approach 1:
The patent divides the compression domain into multiple compression units (e.g., blocks or chunks of data) that can be independently compressed and managed. When data changes, only the affected compression unit needs to be decompressed and recompressed, rather than the entire compression domain. This segmentation resolves the contradiction by enabling efficient updates while maintaining reasonable compression ratios through localized processing.
2Loss of substance
If the compression domain is made larger to improve compression efficiency, then more repeating sequences can be found, but the decompress-modify-recompress operations become more resource-intensive
Solution Approach 1:
By segmenting the compression domain into smaller compression units, the patent reduces the amount of data that needs to be decompressed and recompressed during modifications. This allows the system to maintain good compression ratios (by still being able to find repeating sequences within each unit) while significantly reducing the computational resources required for update operations.
Solution Approach 2:
The patent applies compression selectively to portions of data rather than always compressing the entire domain. By compressing only when and where it provides benefit (e.g., when repeating sequences are detected), the system achieves good compression ratios without the excessive resource consumption that would result from always compressing large domains.
3Ease of operation
If file-based compression is used where the file construct is the boundary of the compression domain, then compression can be managed at a higher level, but it cannot be implemented by low level storage devices and requires additional meta-data management
Solution Approach 1:
The patent segments the compression domain into fixed-size blocks or chunks that can be independently managed. This segmentation allows compression to be implemented at the storage device level (low-level) without requiring file-system awareness, while still providing structured management capabilities. Each block can be independently compressed, tracked, and updated, reducing the complexity overhead compared to file-based approaches.
Data Source
AI summary
Systems, methods, and computer readable media for lazy compression of data incoming to a data storage entity are disclosed. According to one aspect, a method for lazy compression of data incoming to a data storage entity includes defining at least a portion of the data storage area within the data storage entity as a compressed logical unit for storing at least some data in compressed form; receiving a command to write data to the compressed logical unit, and, in response to receiving the command to write data to the compressed logical unit, writing the data in uncompressed form; and monitoring for a trigger condition, wherein, upon detection of a trigger condition, at least a portion of the uncompressed data within the compressed logical unit is compressed.