Tiered Flash Storage Allocation by Write Activity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems using Flash-based storage devices face inefficiencies due to differences in write endurance and speed among various Flash technologies, leading to suboptimal allocation of data objects based on their activity levels.
Innovation Solution
A method and system that categorize data objects into tiers based on write-activity levels, using high-write-endurance devices like MLC Flash for active data and lower-write-endurance devices like TLC Flash for less active data, with metadata management to optimize storage and reduce wear on devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data objects are stored in high-write-endurance devices (MLC Flash), then write reliability and endurance are improved, but storage cost and capacity are increased
Solution Approach 1:
The storage system is segmented into multiple tiers with different Flash memory types (MLC, TLC, QLC) based on write endurance characteristics. Data objects are divided and placed in appropriate tiers based on their write-activity levels, allowing each storage tier to be optimized for its specific function rather than using a uniform storage approach throughout the system.
Solution Approach 2:
Different storage locations are assigned different qualities (write endurance characteristics) matched to the specific needs of data objects. Frequently written data objects are placed in MLC Flash with high write endurance, while less frequently written data objects are placed in TLC or QLC Flash with lower write endurance but higher capacity, creating local optimization rather than global uniformity.
2Speed
If data objects are stored in high-write-endurance devices (MLC Flash), then write speed and performance are improved, but storage cost is increased
Solution Approach 1:
The storage system segments data objects based on their write-activity levels and places them in appropriate storage tiers. Hot data with high write activity is stored in high-performance MLC Flash for fast writes, while cold data with low write activity is stored in lower-cost TLC or QLC Flash, achieving performance optimization without uniformly incurring high costs across all storage.
Solution Approach 2:
The system dynamically changes the storage parameters (location tier) of data objects based on their write-activity level. By monitoring write patterns and adjusting storage location accordingly, the system adapts performance and cost characteristics to match actual usage patterns, achieving cost-effective performance optimization.
3Quantity of substance
If data objects are stored in lower-write-endurance devices (TLC Flash), then storage capacity and cost efficiency are improved, but write reliability and durability are reduced
Solution Approach 1:
Lower-write-endurance storage devices (TLC, QLC Flash) are assigned to store data objects with low write-activity levels, creating a local match between storage capability and data requirements. This ensures that the limitations of lower-endurance devices are not exposed to workloads that would exceed their capacity, while their high capacity benefits are fully utilized for appropriate data types.
Solution Approach 2:
The patent converts the potential harm of using lower-write-endurance devices into a benefit by strategically placing only suitable data objects in those devices. The lower write endurance becomes irrelevant for cold data that rarely requires writing, while the higher capacity and lower cost become the dominant advantages, effectively transforming a limitation into an asset.
4Device complexity
If all data objects are stored in the same storage devices, then device complexity is reduced, but storage efficiency and resource utilization are worsened
Solution Approach 1:
The storage system implements dynamic data placement and migration between tiers based on changing write-activity levels. Data objects can move from hot tiers to cold tiers or vice versa as their access patterns evolve, allowing the system to adapt to dynamic workloads and maintain high storage efficiency without requiring static, overly complex manual management.
Solution Approach 2:
The storage system automatically monitors write-activity levels and performs data migration decisions without requiring complex external intervention. The tiering mechanism self-adjusts based on observed usage patterns, reducing the need for complex external storage management while maintaining high efficiency through automated optimization.
Data Source
AI summary
A method includes receiving data objects for storage in at least one or more first storage devices characterized by a first write endurance, and one or more second storage devices characterized by a second write endurance that is lower than the first write endurance. Respective write-activity levels are estimated for the data objects. For each data object, a choice is made whether to store the data object in the first storage devices or in the second storage devices based on an estimated write-activity level of the data object. The data object is stored in the chosen storage devices.