Dynamic Storage Tier Migration via Variable Slice Sizing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data storage systems lack efficient methods for dynamically managing data placement across different storage tiers based on I/O workload statistics, leading to suboptimal performance and resource utilization.

Innovation Solution

A method that involves receiving data portions at logical addresses and storing them on multiple slices of physical storage across various tiers, with dynamic relocation based on I/O workload statistics, including promotion and demotion between higher and lower performance tiers, and adjusting slice sizes to optimize storage and performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If data is stored on fixed-size slices across all storage tiers, then storage allocation is simplified, but storage efficiency and performance optimization are reduced

Engineering Contradiction:

Improvestorage allocation simplicityVSAvoidstorage efficiency

Core Design Contradiction:

Ease of manufactureVSProductivity

Solution Approach 1:

The patent segments storage slices into different sizes based on storage tier characteristics. Instead of using uniform slice sizes across all tiers, the system creates variable-sized slices (e.g., 64KB, 128KB, 256KB) that are optimized for specific tier requirements, allowing high-performance tiers to use smaller slices for faster access while capacity-oriented tiers use larger slices for better utilization

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the size parameter of storage slices dynamically based on the storage tier. Each storage tier is assigned a specific slice size parameter that optimizes performance for that tier's characteristics, enabling the system to adapt slice dimensions to match the performance-critical or capacity-critical nature of different tiers

Inventive Principle:

Principle #35Parameter changes

2Productivity

If data relocation between storage tiers is performed frequently to optimize performance, then system performance improves, but overhead and complexity increase

Engineering Contradiction:

Improvesystem performanceVSAvoiddata management complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements feedback mechanisms that monitor I/O workload statistics and performance metrics of data on each storage tier. Based on this feedback, the system automatically triggers data relocation when performance thresholds are met or exceeded, enabling dynamic optimization without manual intervention while reducing unnecessary relocations through threshold-based control

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent makes the data relocation process dynamic by allowing slice sizes and relocation thresholds to be adjusted based on changing workload patterns. The system adapts relocation behavior in real-time based on observed I/O patterns, promoting or demoting data between tiers dynamically rather than following fixed schedules

Inventive Principle:

Principle #15Dynamics

3Ease of operation

If uniform slice sizes are used across all storage tiers, then storage management is simplified, but performance optimization for different workload types is reduced

Engineering Contradiction:

Improvestorage management simplicityVSAvoidworkload optimization capability

Core Design Contradiction:

Ease of operationVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by assigning different slice sizes to different storage tiers based on their specific requirements. High-performance tiers (e.g., SSDs) use smaller slice sizes for faster access times, while capacity-oriented tiers (e.g., HDDs) use larger slice sizes for better space utilization, allowing each tier to be optimized for its specific function

Inventive Principle:

Principle #3Local quality

4Speed

If data is kept on high-performance storage tiers, then access speed improves, but storage resource utilization and cost efficiency decrease

Engineering Contradiction:

Improvedata access speedVSAvoidstorage resource utilization

Core Design Contradiction:

SpeedVSQuantity of substance

Solution Approach 1:

The patent implements periodic monitoring and evaluation of data access patterns to determine when data should be promoted to or demoted from high-performance tiers. By periodically assessing I/O workload statistics and performance metrics, the system ensures data resides on optimal tiers based on current usage patterns, balancing speed requirements with resource utilization efficiency

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS11467751B2Migation of data portions between different sized slices in the same of different performance tier

Publication Date: 2022.10.11 EMC IP HLDG CO LLC

AI summary

Techniques for data placement may include receiving data portions stored at logical addresses, and storing the data portions on slices of physical storage located in storage tiers. The storage tiers may include different size slices on the different tiers. In one embodiment, slices in the same tier are all the same size. In another embodiment, slices in the same tier may be of different sizes. Slices of data may be demoted and promoted among the storage tiers as the workloads of the slices changes over time. Demotion may include combining slices into a larger slice. Promotion may include partitioning a slice into smaller slices. Additionally, multiple slices of a tier may be combined into a larger slice in the tier. A slice in the tier may be partitioned into multiple smaller slices also located in the tier.