Incremental Memory Dump with Parallel Reboot

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional systems for dumping main memory content and CPU states in large virtual machines or logical partitions take several hours, leading to significant system outages and requiring twice the original memory, with the inability to boot during the dump process.

Innovation Solution

A system that manages a pre-defined set of minimal dedicated resources for executing the dump and reboot, allowing the process to start with a minimum resource availability, and continuously allocates additional resources until the full configuration is reached, enabling parallel execution of dumping and rebooting.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional dumping process is used to store memory content and CPU states, then complete system dump is achieved, but system outage time increases significantly

Engineering Contradiction:
Improvedump completenessVSAvoidsystem outage time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by starting the reboot process before the dump process is fully completed. The system initiates reboot with available dumped data, then continues dumping remaining memory content in the background. This allows the system to begin recovery operations while still gathering complete diagnostic information, thereby reducing overall outage time without sacrificing dump completeness.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamics by making the resource allocation flexible during the dump and reboot process. Instead of requiring all resources to be available before starting reboot, the system dynamically allocates resources as they become available during dumping. The reboot process can start with partial resources and acquire additional resources during execution, enabling parallel operation of dump and reboot activities.

Inventive Principle:
Principle #15Dynamics

2Reliability

If traditional dumping process is used to store all memory content, then complete dump is achieved, but memory resource requirements double

Engineering Contradiction:
Improvedump completenessVSAvoidmemory resource
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent applies segmentation by dividing the memory dumping process into multiple segments or phases. Instead of requiring all memory to be dumped simultaneously into available space, the system dumps memory in segments as they become available. This allows the dump process to proceed with smaller memory allocations at any given time, reducing the peak memory resource requirement from double the original to a more manageable amount while still achieving complete dump coverage.

Inventive Principle:
Principle #1Segmentation

3Reliability

If traditional system waits for complete dump before reboot, then full dump data is available, but resource utilization efficiency decreases

Engineering Contradiction:
Improvedump data availabilityVSAvoidresource utilization efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent implements continuity of useful action by ensuring that both dump and reboot processes continue simultaneously rather than sequentially. The dump process continues in the background while the reboot process progresses, maximizing resource utilization. Resources that would otherwise remain idle during a sequential approach are actively used for both dumping and rebooting in parallel, thereby improving resource utilization efficiency while maintaining complete data availability.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentUS10606681B2Incremental dump with fast reboot
Publication Date: 2020.03.31 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US10606681B2 patent drawing
  • US10606681B2 patent drawing
  • US10606681B2 patent drawing

AI summary

Methods and systems for executing dumping of main memory content and CPU states and for an adaptive boot. The methods and the systems provide a configuration list of the computer system comprising a pre-defined set of dedicated resources for the dumping, provide threshold values for a pre-defined set of minimum resources for executing a reboot of the computer system, assign the pre-defined set of the dedicated resources for executing the dumping, start the dumping, release ones of the dedicated resources after content of the ones of the dedicated resources has been dumped, start a reboot process of the computer system in response to determining that the ones of the dedicated resources exceeds the threshold values for the pre-defined set of the minimum resources for executing the reboot process, and continue to release others of the dedicated resources to the reboot process until the dumping is completed.