Distributed Control Resource Failure Mitigation Through State Checkpointing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Resource failures in distributed control systems (DCS) compromise system stability and availability, leading to service interruptions, data loss, and potential performance degradation due to unmanaged workload redistribution.

Innovation Solution

A method and system for predicting resource failure by analyzing diagnostic data, persisting the service state through checkpointing, and restoring the service on a second resource using the persisted state, utilizing container checkpointing and load balancing to minimize downtime and data loss.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If redundancy is designed into the DCS, then system resiliency is improved, but device complexity and cost increase

Engineering Contradiction:
Improvesystem resiliencyVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by predicting resource failures before they occur using diagnostic data analysis. When failure is predicted, the service state is persisted and the service is restored on a different resource proactively, rather than reactively after failure. This preliminary action ensures system resiliency without requiring complex redundant hardware configurations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system creates a copy of the service state through checkpointing before failure occurs. This copied state is then restored on a different resource, enabling service continuity without requiring identical redundant hardware. The copying approach simplifies the system architecture compared to full hardware redundancy.

Inventive Principle:
Principle #26Copying

2Reliability

If workload is redistributed to other resources, then service continuity is maintained, but performance degradation occurs due to excessive load

Engineering Contradiction:
Improveservice continuityVSAvoidsystem performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system performs workload redistribution in advance when failure is predicted, rather than forcing immediate redistribution after failure. This allows the target resource to be properly prepared and selected, avoiding excessive load concentration. The preliminary action includes selecting an appropriate target resource and smoothly transitioning the workload.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses diagnostic data and performance metrics as feedback to monitor resource health and predict failures. This feedback mechanism enables proactive workload management, allowing the system to redistribute load before failure occurs and preventing performance degradation by avoiding forced redistribution under stress conditions.

Inventive Principle:
Principle #23Feedback

3Reliability

If checkpointing is performed frequently, then data loss risk is reduced, but system overhead and performance impact increase

Engineering Contradiction:
Improvedata loss riskVSAvoidsystem overhead
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system performs checkpointing in advance when failure is predicted, rather than continuously or at fixed intervals. This on-demand checkpointing approach captures the service state only when necessary, minimizing overhead while ensuring data safety. The preliminary checkpointing occurs at an optimal moment before failure, reducing both data loss risk and system overhead.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4607353A1Resource failure mitigation
Publication Date: 2025.08.27 ABB (SCHWEIZ) AG
  • EP4607353A1 patent drawingFigure 1
  • EP4607353A1 patent drawingFigure 2
  • EP4607353A1 patent drawing

AI summary

Resource failure in a distributed control system (DCS) can compromise stability and availability of the system. There is therefore provided a resource failure mitigation method for a distributed control system. The method comprises: predicting failure of a first resource executing a service; persisting a state of the service; and restoring the service at a second resource using the persisted state. A corresponding system is also provided. Risk of data loss can thereby be reduced while improving resiliency, resulting in an improved user experience for non-critical services.