Storage Data Segmentation for Parallel Tier Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face challenges in efficiently managing data across multiple tiers of storage with varying performance characteristics and costs, as current techniques do not scale well with increasing storage amounts and fail to react quickly to workload changes.
Innovation Solution
The system subdivides data into portions, determines independent scores based on access metrics, and dynamically places sub-portions into appropriate storage classes, using short-term and long-term access statistics for promotion and demotion, respectively, to optimize data placement across different storage tiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data is completely analyzed before identifying candidates for promotion and demotion between tiers, then placement accuracy is improved, but processing time increases and scalability deteriorates
Solution Approach 1:
The patent segments the data analysis process into independent score calculations for each data portion, allowing parallel processing. Each data portion receives an independent score based on access metrics, eliminating the need to wait for complete analysis of all data before making placement decisions. This segmentation enables continuous, parallel data movement between tiers while maintaining placement accuracy through individualized scoring.
2Measurement precision
If small region size is used for tier selection, then optimization precision is improved, but the number of regions to process increases significantly
Solution Approach 1:
The patent applies segmentation by dividing data into portions of varying sizes and calculating independent scores for each. This allows the system to maintain fine-grained optimization precision while managing complexity through parallel processing of multiple portions. The independent scoring mechanism enables efficient handling of large numbers of regions without sequential processing bottlenecks.
3Ease of operation
If sequential mode of operation is used for data analysis, then processing simplicity is maintained, but responsiveness to workload changes deteriorates
Solution Approach 1:
The patent introduces dynamics by enabling parallel processing of data portions with independent score calculations. This dynamic approach allows the system to respond quickly to workload changes while maintaining operational simplicity through automated continuous monitoring and placement adjustments. The parallel processing capability enables real-time adaptation to changing access patterns without complex manual intervention.
4Adaptability or versatility
If continuous data movement between tiers is implemented, then system adaptability is improved, but system stability during data movement deteriorates
Solution Approach 1:
The patent segments data movement into parallel operations on independent data portions, allowing continuous adaptability while maintaining stability. By processing portions independently and in parallel, the system can continuously adjust data placement without disrupting overall system stability. The segmented approach enables multiple data portions to be moved simultaneously without creating instability from sequential reconfiguration.
Data Source
AI summary
Identifying data for placement in a storage system having a plurality of storage classes includes subdividing the data into portions, for each of the portions, independently determining at least one score for a particular portion based on a metric corresponding to access of the particular portion, where the at least one score for the particular portion is independent of scores for other ones of the portions, and identifying sub-portions of data for placement in a particular storage class based on the at least one score of a portion of data corresponding to the sub-portions. The at least one score may be based on short term access statistics and long term access statistics. The access statistics may include read misses, writes, and prefetches.