Storage Device AI Processor Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional storage devices lack an AI function, leading to inefficient power consumption and increased data traffic when performing AI calculations, as they rely on resources and accelerators in the host device, which can result in bottlenecks and inefficiencies.

Innovation Solution

A storage device with a separate second processor and nonvolatile memory for AI calculations, allowing it to perform AI functions independently of the host device, reducing data traffic and power consumption by only exchanging input and result data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI calculations are performed using host device resources and accelerators, then AI processing capability is provided, but power consumption increases and data traffic bottlenecks occur

Engineering Contradiction:
ImproveAI processing capabilityVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The system is divided into two independent processing domains: the host device for general operations and the storage device with its own processor for AI calculations. This segmentation allows AI processing to occur independently within the storage device, eliminating the need to consume host device power resources while maintaining AI processing capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The storage device acts as an intermediary between the host device and the AI processing function. By embedding a processor within the storage device, it serves as a localized AI processing unit that reduces data traffic to the host while providing AI capabilities, thus resolving the power consumption and data traffic bottleneck issues.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If AI calculations are performed using host device resources, then AI processing capability is provided, but data traffic bottlenecks occur

Engineering Contradiction:
ImproveAI processing capabilityVSAvoiddata traffic
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The system is divided into two independent processing domains: the host device for general operations and the storage device with its own processor for AI calculations. This segmentation allows AI processing to occur independently within the storage device, eliminating the need to consume host device power resources while maintaining AI processing capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The storage device acts as an intermediary between the host device and the AI processing function. By embedding a processor within the storage device, it serves as a localized AI processing unit that reduces data traffic to the host while providing AI capabilities, thus resolving the power consumption and data traffic bottleneck issues.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS11468306B2Storage device with artificial intelligence and storage system including the same
Publication Date: 2022.10.11 SAMSUNG ELECTRONICS CO LTD
  • US11468306B2 patent drawing
  • US11468306B2 patent drawing
  • US11468306B2 patent drawing

AI summary

A storage system includes a host device and a storage device. The host device provides first input data for data storage function and second input data for artificial intelligence (AI) function. The storage device stores the first input data from the host device, and performs AI calculation based on the second input data to generate calculation result data. The storage device includes a first processor, a first nonvolatile memory, a second processor and a second nonvolatile memory. The first processor controls an operation of the storage device. The first nonvolatile memory stores the first input data. The second processor performs the AI calculation, and is distinguished from the first processor. The second nonvolatile memory stores weight data associated with the AI calculation, and is distinguished from the first nonvolatile memory.