Storage Array Compression Using Activity and Compressibility Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current naive activity-based compression techniques in storage arrays inefficiently consume resources by compressing data with low or no reducibility, degrading performance while trying to meet data reduction requirements.
Innovation Solution
Implementing activity and compressibility-based data compression that dynamically identifies and prioritizes address spaces with specific activity and compressibility characteristics to maintain or enhance storage array performance, using a controller to analyze IO workloads, generate activity and compressibility models, and adjust compression thresholds based on observed data reduction requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If activity-based compression is applied to all address spaces, then data reduction requirements are met, but performance degrades due to compressing low-compressibility data
Solution Approach 1:
The patent applies different compression strategies to different address spaces based on their local characteristics. Activity-based compression is applied only to address spaces exhibiting specific activity patterns (e.g., sequential writes, cold data), while other address spaces are excluded from compression. This selective approach ensures that compression resources are focused on data that will actually benefit from it, maintaining performance while achieving data reduction goals.
Solution Approach 2:
The system dynamically adjusts compression eligibility based on observed activity patterns. Address spaces are monitored for activity characteristics, and compression status is adjusted in response to changing workloads. This dynamic adaptation allows the system to optimize between performance and data reduction based on actual usage patterns rather than static rules.
2Quantity of substance
If compression is applied to meet data reduction requirements, then storage capacity increases, but resource consumption increases and performance degrades
Solution Approach 1:
Instead of applying compression universally or excessively to meet data reduction targets, the patent applies compression partially and selectively only to eligible address spaces. This partial action approach achieves sufficient data reduction by focusing on high-compressibility data while avoiding the excessive resource consumption that would result from compressing all data regardless of compressibility characteristics.
3Loss of substance
If compression thresholds are lowered to increase data reduction, then more data is compressed, but performance degradation increases
Solution Approach 1:
The patent changes the selection parameters for compression by using activity-based criteria instead of purely compressibility-based or LBA-based criteria. By monitoring activity patterns (such as sequential write patterns, access frequency, or data age), the system identifies address spaces that are good compression candidates based on their behavioral characteristics, achieving effective data reduction without the performance penalty of aggressive compression thresholds.
Data Source
AI summary
One or more aspects of the present disclosure relate to increasing the performance of a storage array using activity and compressibility-based data compression. In embodiments, an input/output (IO) workload is received at a storage array. Additionally, at least one address space of at least one logical storage device with an IO activity corresponding to an activity characteristic and compressibility corresponding to a compressibility characteristic is identified. Further, if a data reduction requirement (DRR) of the storage array is unsatisfied, IO write requests of the IO workload targeting the identified address spaces of each logical device are compressed.