SMR Drive Sequential Write Regions and Buffer Mapping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Shingled magnetic recording (SMR) drives in cloud data storage systems experience significant latency and write amplification due to their sequential writing nature, which is problematic for random write operations, impacting overall system performance.

Innovation Solution

The system divides the magnetic storage device into multiple sequential write regions with guard spaces and uses a non-volatile buffer to store key-value pairs, where new versions are written in separate slabs, with a mapping table tracking locations, allowing for efficient sequential writing and reducing latency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If SMR drives are used to increase storage density, then storage capacity is improved, but write latency increases significantly

Engineering Contradiction:
Improvestorage capacityVSAvoidwrite latency
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-allocating sequential write regions and guard bands on SMR drives before random write operations occur. The system divides the storage device into multiple sequential write regions with guard spaces, allowing random writes to be redirected to appropriate sequential regions without causing track overwrite conflicts. This preprocessing of the storage structure enables subsequent random writes to proceed with minimal latency while maintaining SMR's high storage density capability.

Inventive Principle:
Principle #10Preliminary action

2Quantity of substance

If sequential writing is used on SMR drives, then storage density is improved, but write amplification increases

Engineering Contradiction:
Improvestorage densityVSAvoidwrite amplification
Core Design Contradiction:
Quantity of substanceVSLoss of substance

Solution Approach 1:

The patent applies segmentation by dividing the SMR drive into multiple independent sequential write regions, each with its own guard band. This segmentation allows the system to perform targeted sequential writes to specific regions rather than requiring full-band sequential writes. By localizing write operations to only the necessary segments, the patent reduces write amplification while maintaining the storage density benefits of SMR technology.

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If random write operations are performed on SMR drives, then data accessibility is improved, but performance deteriorates due to latency and write amplification

Engineering Contradiction:
Improvedata accessibilityVSAvoidwrite performance
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent introduces an intermediary layer (the sequential write region mapping system) between the random write requests and the physical SMR media. This intermediary translates random write operations into sequential writes within appropriate regions, using guard bands to prevent track overwrites. The mapping system acts as a mediator that preserves the ease of random data accessibility while converting operations to maintain high write performance on SMR drives.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS11093149B2Method to efficiently store object data of an object storage service on a magnetic disk drive and magnetic SMR disk drive
Publication Date: 2021.08.17 KIOXIA CORP
  • US11093149B2 patent drawing
  • US11093149B2 patent drawing
  • US11093149B2 patent drawing

AI summary

In a data storage system, the available space of a magnetic storage device is divided into multiple sequential write regions for storing sequentially written data, where the regions are each separated from adjacent sequential write regions by a guard space. Object data, such as key-value pairs, are written sequentially to a particular sequential write region, in blocks of data that correspond to the contents of a nonvolatile buffer being flushed to the magnetic storage device. When a key-value pair stored in the magnetic storage device is subsequently updated, the original key-value pair is not overwritten. Instead, the new version of the key-value pair is included in the next block of data to be written to the magnetic storage device, and a mapping table tracks the location of the newest version of each key-value pair stored in the magnetic storage device.