SSD Cache RMW Offload via Header Linking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Read-modify-write (RMW) operations in data storage systems consume more SSD cache space, conflicting with the desire to maximize the use of SSD devices as caching mediums due to the need to store modified data before persistence, which affects I/O performance.

Innovation Solution

Offloading RMW operations to an SSD cache layer, where new and old RMW headers are managed within a caching medium, allowing for efficient data management and reduced SSD cache space usage by writing new data directly to the cache and maintaining links between headers for efficient read and flush operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If RMW operations are performed on the logical volume, then data integrity is maintained, but SSD cache space consumption increases

Engineering Contradiction:
Improvedata integrityVSAvoidSSD cache space
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the RMW operation into two distinct phases: (1) writing new data to the SSD cache layer, and (2) separately managing old data through header links. This segmentation allows the system to maintain data integrity while reducing the need to hold entire modified chunks in cache simultaneously, thereby reducing SSD cache space consumption.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces RMW headers as intermediary structures that manage the relationship between new and old data. These headers act as mediators that track data modifications without requiring the entire old data to be retained in the SSD cache, thus maintaining reliability while optimizing cache space utilization.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Stability of the object's composition

If modified data is stored in SSD cache before persistence, then data consistency is ensured, but cache utilization efficiency decreases

Engineering Contradiction:
Improvedata consistencyVSAvoidcache utilization efficiency
Core Design Contradiction:
Stability of the object's compositionVSProductivity

Solution Approach 1:

The patent performs preliminary actions by writing new data to the SSD cache layer immediately upon receiving write I/O operations, before the flush operation is triggered. This preliminary action ensures data consistency is maintained while allowing the system to defer the more space-intensive operations (reading old data, merging, and writing to logical volume) until later, thereby improving cache utilization efficiency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamic data management where the treatment of data in the SSD cache changes based on the flush operation status. New data is written dynamically to the cache, and the system dynamically determines which old data needs to be retained based on the specific flush requirements, optimizing both consistency and efficiency.

Inventive Principle:
Principle #15Dynamics

3Reliability

If entire chunk data is read and modified for sub-chunk writes, then data integrity is maintained, but I/O operations increase

Engineering Contradiction:
Improvedata integrityVSAvoidI/O performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies local quality by treating new data and old data differently in terms of storage location and management. New data is written to the SSD cache layer with local address translation, while old data is managed through header links. This localized differentiation allows the system to maintain data integrity for the entire chunk while performing I/O operations only on the specific sub-chunk that needs modification, thereby improving I/O performance.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS10114566B1Systems, devices and methods using a solid state device as a caching medium with a read-modify-write offload algorithm to assist snapshots
Publication Date: 2018.10.30 AMZETTA TECH LLC
  • US10114566B1 patent drawing
  • US10114566B1 patent drawing
  • US10114566B1 patent drawing

AI summary

A method for offloading RMW operations on a volume of a storage system to a caching medium includes providing an SSD cache, a data structure for managing the SSD cache and cache headers for managing cache lines. The method also includes maintaining cache headers for cache lines corresponding to a same chunk of the volume as a group, creating a snapshot of the volume and, after creating the snapshot, receiving a write I/O operation directed to a portion of the same chunk. The method further includes allocating a new RMW header for the write I/O operation, maintaining the new RMW header in a new group, performing the write I/O operation on the SSD cache, allocating an old RMW header for the group, maintaining the old RMW header with the group and providing a link between the new and old RMW headers.