Memory Read Buffering for Repeated Target Address Requests

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Memory systems face inefficiencies in processing multiple read commands for the same target address, leading to prolonged processing times due to repeated access to memory media.

Innovation Solution

Implementing a controller with a buffer to store read data and generate response messages based on buffered data, managing read commands through a link list to avoid repeated memory access, and suspending write operations until all read commands are processed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If the memory system processes each read command by accessing the memory media separately, then each command is handled individually, but the processing time increases due to repeated memory access

Engineering Contradiction:
Improveprocessing timeVSAvoidcommand processing efficiency
Core Design Contradiction:
Loss of timeVSProductivity

Solution Approach 1:

The patent applies preliminary action by reading and caching data from the memory media into a buffer before receiving subsequent read commands for the same address. When the controller receives a read command, it first checks the buffer to see if the requested data is already cached. If found, it serves the data from the buffer without accessing the memory media again, thereby eliminating redundant access operations and reducing processing time for repeated read commands.

Inventive Principle:
Principle #10Preliminary action

2Speed

If the controller uses a buffer to cache read data and avoid repeated memory access, then processing speed improves, but the device complexity increases

Engineering Contradiction:
Improveprocessing speedVSAvoidcontroller structure
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent introduces a buffer as an intermediary component between the memory media and the controller's command processing logic. This buffer serves as a caching layer that stores recently accessed data, allowing the controller to serve subsequent read requests without directly accessing the memory media. The buffer acts as a mediator that absorbs repeated read requests and satisfies them from cached data, thereby improving processing speed while adding only a single buffering component rather than complex reconfiguration of the entire system.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If the memory system processes multiple read commands for the same address simultaneously, then throughput increases, but data consistency may be compromised during write operations

Engineering Contradiction:
Improvecommand throughputVSAvoiddata consistency
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the controller monitors incoming commands and maintains a record of pending read commands for each memory address. When a write command is received, the controller checks this record to identify any outstanding read commands for the same address. The controller then waits for these read commands to be served from the buffer before executing the write operation, ensuring that read operations see consistent data. After the write is completed, the buffer is updated and the read command records are cleared, maintaining data consistency while allowing high throughput for read-heavy workloads.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20260050398A1Memory system and method for processing read commands for a target address
Publication Date: 2026.02.19 SK HYNIX INC
  • US20260050398A1 patent drawing
  • US20260050398A1 patent drawing
  • US20260050398A1 patent drawing

AI summary

A memory system may include a memory media and a controller. The controller may receive from a host a first read command which requests to read target data stored in a target address of the memory media, read the target data from the memory media, store the read target data in a buffer, generate a first response message to the first read command on the basis of the read target data, receive, after receiving the first read command, from the host a second read command which requests to read the target data, and generate a second response message to the second read command on the basis of the target data stored in the buffer.