Scale-Out Storage Controller Data Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flash memory devices' performance is degraded when one of the multiple flash memory chips wears out, leading to a decrease in overall data storage device lifetime, as existing technologies do not effectively manage and distribute data based on the characteristics of individual memory chips within a scale-out storage system.
Innovation Solution
A data storage device with a controller that determines the characteristics of each scale-out storage device, including volatile and nonvolatile memory, and transmits data accordingly, using a type formatter and classifier to optimize data distribution based on factors like operation frequency, bandwidth, and memory cell programming methods, ensuring efficient data management across multiple scale-out storage devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is stored in multiple flash memory chips, then storage capacity is increased, but overall device lifetime is degraded when one chip wears out
Solution Approach 1:
The patent segments the storage system into multiple independent scale-out storage devices, each with its own controller and memory chips. The controller classifies data characteristics and distributes different types of data to different storage devices based on their individual performance characteristics. This segmentation allows the system to treat each chip independently, so wear in one chip does not necessarily degrade the performance of other chips, thereby resolving the contradiction between increased storage capacity and maintained device lifetime.
Solution Approach 2:
The patent applies local quality by characterizing each scale-out storage device's memory chips individually and assigning them specific roles based on their unique performance characteristics. The controller determines the type of each storage device and classifies data accordingly, placing data with appropriate access patterns and durability requirements on chips that best match those requirements. This ensures that each chip operates within its optimal performance envelope, extending overall system lifetime while maintaining total storage capacity.
2Productivity
If data is uniformly distributed across all scale-out storage devices, then storage utilization is maximized, but performance is degraded due to varying chip characteristics
Solution Approach 1:
The patent changes the distribution parameter from uniform to characteristic-based allocation. The controller determines the type of each scale-out storage device by evaluating their performance parameters (such as read/write speeds, endurance, latency) and dynamically classifies data characteristics (hot, warm, cold data). Data is then distributed to storage devices whose characteristics match the data requirements, maximizing both utilization and performance by avoiding placement of performance-sensitive data on degraded chips.
Solution Approach 2:
The patent introduces dynamic adaptability through the type formatter and classifier components that continuously evaluate storage device characteristics and data access patterns. The system can dynamically adjust data distribution strategies based on real-time performance monitoring, migrating data between storage devices as chip characteristics change over time due to wear. This dynamic approach maintains optimal performance while fully utilizing available storage capacity across all devices.
3Reliability
If the system monitors and manages each memory chip individually, then device lifetime is extended, but system complexity increases
Solution Approach 1:
The patent implements a universal controller architecture that handles multiple functions: determining storage device types, classifying data characteristics, making distribution decisions, and monitoring chip health. This multi-functional controller consolidates what would otherwise require separate management systems for each chip, extending device lifetime through comprehensive monitoring while managing complexity through a unified control plane that operates across all scale-out storage devices.
4Reliability
If data is classified and distributed based on storage device characteristics, then performance is optimized, but data management complexity increases
Solution Approach 1:
The patent applies preliminary action by having the type formatter pre-determine the characteristics of each scale-out storage device and store this information for future reference. The classifier is pre-configured with data characteristic categories (hot, warm, cold). When data needs to be stored or accessed, the system quickly matches data characteristics against pre-established storage device profiles, avoiding complex real-time analysis. This preliminary classification and profiling significantly reduces the complexity of ongoing data management operations while maintaining optimized performance.
Data Source
AI summary
A data storage device includes a plurality of scale-out storage devices, and a controller configured to determine a characteristic of data and configured to transmit and receive the data to or from a first scale-out storage device among the plurality of scale-out storage devices based on the determined characteristic. Each of the plurality of scale-out storage devices includes a volatile memory, a nonvolatile memory, and a scale-out controller configured to control the volatile memory and the nonvolatile memory.