Shared Memory Management via Access Blocking and Data Copying

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In information processing systems with multiple connected devices, when a failure occurs in a shared memory, existing technologies fail to ensure that written data is correctly transferred to a replacement memory, leading to potential data loss and system instability due to the inability to prevent access by other operating systems.

Innovation Solution

A shared memory management method that uses memory tokens and access tokens to control access, detects errors, and performs a replacing process by identifying and blocking access to the failed segment, copying data to a replacement segment, and updating management tables to ensure data integrity and system continuity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the automatic replacing function is used to copy data from a failed memory area to a normal memory area, then data can be recovered, but data loss occurs when another information processing device writes data to the failed memory area during the copying process

Engineering Contradiction:
Improvedata recoveryVSAvoiddata loss
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent applies preliminary anti-action by having the detecting unit detect memory errors and the preventing control unit block access to the failed memory area before the automatic replacing function can be compromised. This prevents other information processing devices from writing data to the failed area during the copying process, thereby preventing the data loss that would otherwise occur

Inventive Principle:
Principle #9Preliminary anti-action

Solution Approach 2:

The patent implements preliminary action by having the preventing control unit establish access restrictions on the failed memory area in advance of the data copying operation. This ensures that when the automatic replacing function copies data from the failed area to a normal area, no other device can interfere by writing new data to the failed area, thus guaranteeing complete data recovery

Inventive Principle:
Principle #10Preliminary action

2Reliability

If one OS executes the automatic replacing function to prevent data loss, then data integrity is maintained, but access control fails when a different OS writes data to the same memory area

Engineering Contradiction:
Improvedata integrityVSAvoidaccess control
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent applies universality by implementing a cross-OS access control mechanism through the preventing control unit that operates independently of any single operating system. This universal control mechanism can block access to the failed memory area regardless of which OS (OS1 or OS2) is attempting to write data, thereby maintaining data integrity while providing access control that works across multiple operating systems

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent uses an intermediary approach by introducing the preventing control unit as a mediator between multiple operating systems and the shared memory. This intermediary component receives access requests from any OS, checks whether the target memory area has failed, and blocks access if necessary. This mediator enables one OS to execute the automatic replacing function while preventing other OSes from compromising data integrity

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If access to the failed memory area is allowed to maintain system operation, then system continuity is preserved, but data loss occurs when data is written to the failed area during automatic replacing

Engineering Contradiction:
Improvesystem continuityVSAvoiddata loss
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent applies preliminary anti-action by having the detecting unit detect memory errors and the preventing control unit block access to the failed memory area before the automatic replacing function can be compromised. This prevents other information processing devices from writing data to the failed area during the copying process, thereby preventing the data loss that would otherwise occur

Inventive Principle:
Principle #9Preliminary anti-action

Solution Approach 2:

The patent implements the copying principle by having the automatic replacing function create a complete copy of all data from the failed memory area to a normal memory area. Once the copy is complete and verified, the system can safely allow access to the normal area while the failed area is either repaired or isolated. This copying mechanism ensures system continuity while preventing data loss

Inventive Principle:
Principle #26Copying

Data Source

PatentEP2921965B1Information processing device and shared memory management method
Publication Date: 2017.01.11 FUJITSU LTD
  • EP2921965B1 patent drawing
  • EP2921965B1 patent drawing
  • EP2921965B1 patent drawing

AI summary

An access blocking unit (74) blocks an access to a failed segment by using tokens in hardware and a replacing unit (75) performs a process for replacing the failed segment with a replacement segment. For each segment of a shared memory (43), an application recognizing unit (71) recognizes the node numbers of nodes that are given access permission and PIDs of applications and records them in the management table (70). When a failure occurs in the shared memory (43), an access stopping unit (73) identifies applications that use the failed segment including applications of different nodes (1) by using the management table (70) and informs the applications of stop of the use of the failed segment.