Solid-State Storage Multi-Stream Write Collision Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multi-stream data write operations in solid-state storage devices can lead to data write collision, degrading the write speed performance due to garbage collection-induced overhead and workload fluctuations.

Innovation Solution

Implementing a method where data chunks from different write buffers are written in a skewed or randomized sequence across multiple channels and super-pages to reduce collision, utilizing a controller that organizes NAND flash memory chips into channels for parallel access and garbage collection in super-blocks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If multi-stream data write operations are implemented to reduce garbage collection overhead, then garbage collection overhead is reduced, but data write collision increases and write speed performance degrades

Engineering Contradiction:
Improvegarbage collection overheadVSAvoidwrite speed performance
Core Design Contradiction:
Loss of energyVSProductivity

Solution Approach 1:

The patent applies dynamics by making the data write sequence adaptive rather than fixed. The controller dynamically adjusts the write sequence based on real-time channel status and workload conditions, transforming the static multi-stream write approach into a dynamic system that can respond to changing conditions and avoid collisions adaptively.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of write sequence from conventional sequential or fixed patterns to skewed/randomized sequences. By varying the write sequence parameters across different channels and super-pages, the system reduces predictability of collision patterns and improves write speed performance while maintaining garbage collection benefits.

Inventive Principle:
Principle #35Parameter changes

2Stability of the object's composition

If data are written sequentially from multiple write buffers to improve organization, then data organization is improved, but write collision increases due to workload fluctuations

Engineering Contradiction:
Improvedata organizationVSAvoidwrite speed performance
Core Design Contradiction:
Stability of the object's compositionVSReliability

Solution Approach 1:

The patent introduces asymmetry by using skewed sequences instead of symmetric or uniform write patterns. Different channels and write buffers employ different write sequence patterns, creating an asymmetric distribution that reduces the likelihood of simultaneous writes to the same super-page and improves write speed performance.

Inventive Principle:
Principle #4Asymmetry

3Productivity

If parallel data writes across all channels are implemented to improve throughput, then throughput is improved, but data write collision increases

Engineering Contradiction:
ImprovethroughputVSAvoiddata write collision
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent adds another dimension to the write operation by introducing skewed sequences that vary across multiple parameters (channel, super-page, write buffer). This multi-dimensional approach to write distribution reduces collision probability while maintaining parallel throughput across all channels.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10976957B2Reducing multi-stream data write collision in solid-state data storage devices
Publication Date: 2021.04.13 SCALEFLUX INC
  • US10976957B2 patent drawing
  • US10976957B2 patent drawing
  • US10976957B2 patent drawing

AI summary

The present disclosure relates to the field of solid-state data storage, and particularly to improving the speed performance of solid-state data storage devices by reducing multi-stream data write collision. A method according to embodiments includes: writing a plurality of data chunks in a skewed sequence from a plurality of different data write buffers in the storage device into a plurality of different super-pages across all channels of the storage device.