Fault-Tolerant Communications Grid Checkpoint Recovery

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In communications grids, node failures can lead to entire grid failures and job restarts, especially problematic for large data sets or long-running jobs, as existing fault-tolerant mechanisms are inadequate in ensuring seamless recovery.

Innovation Solution

The system detects node failures by generating snapshots of the grid status and project checkpoints, allowing backup nodes to take over and continue project execution from the last known stable state, thereby preventing grid failure and minimizing job restarts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a node fails in a communications grid, then the grid must restart the entire job from the beginning, but this causes significant loss of time and computational resources especially for large data sets or long-running jobs

Engineering Contradiction:
Improvegrid fault toleranceVSAvoidjob restart time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by creating checkpoints and snapshots of grid state at regular intervals during job execution. When a node fails, the system can restore from the most recent checkpoint rather than restarting from scratch, significantly reducing recovery time. This is implemented through the checkpoint creation mechanism that saves grid state periodically during computation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system creates copies of the grid state through snapshots and checkpoints that are stored separately from the active computation. These copies allow the system to restore previous states without losing the entire job progress. The snapshot mechanism creates redundant copies of grid state that can be used for recovery.

Inventive Principle:
Principle #26Copying

2Reliability

If the system implements comprehensive fault detection and recovery mechanisms, then grid reliability improves, but the system complexity increases due to additional monitoring and state management overhead

Engineering Contradiction:
Improvenode failure detection and recoveryVSAvoidgrid monitoring and state management system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system implements self-service fault detection where nodes automatically monitor each other's health status and report failures without requiring external management. The grid state snapshot mechanism automatically detects inconsistencies and triggers recovery procedures without human intervention, reducing the complexity of external monitoring systems.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system uses feedback mechanisms where nodes continuously report their status to the grid management system, which then adjusts its monitoring and recovery actions based on this feedback. The checkpoint and snapshot systems provide feedback about grid state that enables automatic recovery decisions, reducing the need for complex proactive monitoring.

Inventive Principle:
Principle #23Feedback

3Productivity

If the system restores grid state from snapshots after node failure, then job continuity is maintained, but the loss of information occurs due to the gap between the last checkpoint and the actual failure point

Engineering Contradiction:
Improvejob continuityVSAvoidcomputational progress between checkpoints
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The system segments the job execution into discrete checkpoints and snapshots at regular intervals. This segmentation allows the system to recover to the nearest previous checkpoint rather than losing all progress. By dividing the computation into manageable segments, the information loss is limited to only the segment between the last checkpoint and failure, rather than the entire job.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3179368B1Systems and methods for fault tolerant communications
Publication Date: 2018.12.12 SAS INSTITUTE INC
  • EP3179368B1 patent drawingFigure 1
  • EP3179368B1 patent drawingFigure 2
  • EP3179368B1 patent drawingFigure 3

AI summary

Apparatuses, systems and methods are disclosed for tolerating fault in a communications grid. Specifically, various techniques and systems are provided for detecting a fault or failure by a node in a network of computer nodes in a communications grid, adjusting the grid to avoid grid failure, and taking action based on the failure. In an example, a system may include receiving grid status information at a backup control node, the grid status information including a project status, storing the grid status information within the backup control node, receiving a failure communication including an indication that a primary control node has failed, designating the backup control node as a new primary control node, receiving updated grid status information based on the indication that the primary control node has failed, and transmitting a set of instructions based on the updated grid status information.