Hierarchical Storage Data Tier Assignment via Access Frequency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hierarchical storage management systems face challenges in efficiently managing data across different storage tiers based on access frequency, as they often rely solely on past usage patterns, failing to account for historical data access information and specific application requirements, which can lead to suboptimal data placement and performance.
Innovation Solution
A method that determines the overall current data access frequency by combining past and current usage data, allowing for dynamic tier assignment and reassignment based on conditions, while preserving tier assignments for future reference, using heat maps and APIs to manage data placement across storage tiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional HSM systems store bulk data on slower devices and copy data to faster disk drives based on usage monitoring, then storage cost is reduced, but data access performance may be degraded due to insufficient consideration of historical access patterns and application requirements
Solution Approach 1:
The system performs preliminary actions by determining overall current data access frequency before data placement decisions are made. This includes analyzing both past usage patterns and current access behavior to pre-establish optimal tier assignments, ensuring data is proactively positioned in appropriate storage tiers before access demands arise, thus preventing performance degradation while maintaining cost efficiency
Solution Approach 2:
The system implements dynamic data tier assignment by continuously monitoring and re-evaluating data access frequency. Data can be dynamically moved between storage tiers based on changing access patterns, allowing the system to adapt to varying workload requirements and maintain optimal performance while managing storage costs effectively
2Device complexity
If HSM systems rely solely on past usage patterns for data tier assignment, then implementation complexity is reduced, but data placement accuracy deteriorates due to failure to account for current access behavior and application requirements
Solution Approach 1:
The system merges multiple data sources including past usage patterns, current access frequency, and application requirements into a comprehensive data access frequency determination. This combination of historical and real-time information enables more accurate data placement decisions without requiring overly complex individual monitoring systems for each data source
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring current data access frequency and using this information to adjust data tier assignments. The feedback loop allows the system to learn from actual access patterns and refine placement decisions, improving accuracy while maintaining manageable system complexity through iterative optimization
3Speed
If data is frequently moved between storage tiers based on real-time access monitoring, then data access performance is improved, but system operational complexity increases due to continuous monitoring and reassignment operations
Solution Approach 1:
The system employs periodic action by monitoring data access frequency at defined intervals and performing tier reassignment operations based on these periodic evaluations. This approach balances performance optimization with operational simplicity by avoiding continuous real-time monitoring while still capturing significant access pattern changes, thus reducing operational complexity while maintaining improved data access performance
Data Source
AI summary
Embodiments relate to a method for data management. An aspect includes a method for assigning storage types to data based on access frequency. Past or historical data associated with current data usage is also considered prior to assignment. Once data frequency access is determined, the current data is assigned to a first tier of a plurality of hierarchical ordered tiers, each tier corresponding to at least one class of storage. In one embodiment, there may be a condition that overrides the assignment with option to override it. The tier assignment may also be preserved the tier so that the current data can be appropriately assigned in the future.