Non-intrusive Storage Tiering Using Front End Cache Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage devices, discrepancies in data storage management between primary and secondary devices occur due to differences in access frequencies, leading to inefficient data management during failover scenarios, as the secondary device may not be able to support the workload effectively due to differing storage tiering configurations.
Innovation Solution
A method is implemented to determine data demotion candidates based on their presence in the cache and access characteristics, delaying demotion to lower access speed tiers if the data is frequently accessed, using an evaluation algorithm that considers the amount of data in the cache and its residency time, thereby maintaining efficient storage management across multiple storage devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is automatically tiered based on back end access statistics, then storage efficiency is improved, but data that is frequently accessed at the front end may be incorrectly demoted to lower tiers
Solution Approach 1:
The patent merges front end cache access information with back end storage access statistics to make tiering decisions. By combining these two data sources, the system achieves both storage efficiency (from back end statistics) and tiering accuracy (from front end cache information), resolving the contradiction between the two opposing requirements.
Solution Approach 2:
The system implements feedback by monitoring front end cache access patterns and using this information to adjust tiering decisions. The front end provides feedback about actual data access behavior, which corrects potential misclassification from back end statistics alone, ensuring data remains on appropriate performance tiers.
2Loss of energy
If data is demoted to lower access speed tiers, then storage cost is reduced, but performance is degraded for frequently accessed data
Solution Approach 1:
The front end cache access information serves as feedback that prevents demotion of frequently accessed data. This feedback mechanism ensures that data which would benefit from fast access remains on higher performance tiers, avoiding the performance degradation that would result from premature demotion to lower, cheaper tiers.
Solution Approach 2:
The system changes the parameters used for tiering decisions by incorporating front end cache hit ratios and access patterns. This multi-parameter approach allows the system to optimize both cost and performance simultaneously, rather than relying on a single parameter that would force a trade-off between storage cost and access speed.
3Adaptability or versatility
If storage tiering is managed independently at each device, then device autonomy is improved, but data layout consistency across devices deteriorates
Solution Approach 1:
The patent implements a universal tiering policy that operates consistently across multiple storage devices. The same tiering logic and evaluation criteria are applied at each device, ensuring data layout consistency while allowing each device to autonomously execute the policy. This universal approach maintains both device autonomy and cross-device consistency.
Solution Approach 2:
The system uses standardized parameters (such as cache hit ratios, access patterns, and workload characteristics) that are evaluated consistently across all devices. By changing to these universal parameters, each device can independently make tiering decisions that are consistent with other devices in the system, achieving both autonomy and consistency.
Data Source
AI summary
A system and techniques are provided for evaluating front end activity of a storage device in connection with storage tiering and management operations at the back end without the need to collect new sub-LUN metrics/statistics on the front end or querying the front end. The system avoids demotions of extents that are hot on the front end even when seemingly cool on the back end. The result is improved system performance since data that is hot on the front end will not be demoted to a lower storage tier. The system provides that there is little to no additional performance impact on the front end because there is no requirement to collect any new front end statistics or query the front end, and there is no increase in the meta data collected by the automated tiering system as there is no new sub-LUN metrics collected on the front end.