View a markdown version of this page

MIDASUS02-BP01 Actively manage workloads and resource allocation based on production demands - Modern Industrial Data Technology Lens
Implementation guidanceKey AWS servicesResources

MIDASUS02-BP01 Actively manage workloads and resource allocation based on production demands

Identify critical and non-critical manufacturing systems, then align computing resources with actual production schedules and operational requirements to reduce waste while providing reliability for the essential manufacturing systems.

Desired outcome: Cloud resources that efficiently scale with manufacturing operations, prioritizing time-sensitive shop floor systems while optimizing resource usage for enterprise applications, resulting in reduced energy consumption, lower costs, and improved environmental sustainability.

Benefits of establishing this best practice:

  • Reduced energy consumption and carbon footprint by removing over provisioning of resources.

  • Lower operational costs while maintaining reliability for the production systems.

  • Enhanced sustainability reporting metrics with quantifiable improvements in cloud resource efficiency aligned with manufacturing operations.

Level of risk exposed if this best practice is not established: Medium

Implementation guidance

  • Analyze production schedules and system criticality to categorize manufacturing applications as time-critical (shop floor control, real-time monitoring) and non-time-critical (reporting functions, data processing) using workload assessment tools.

  • Implement auto scaling mechanisms for all manufacturing systems based on their specific demand patterns, verifying that enterprise planning and design systems maintain necessary availability while optimizing resource allocation.

  • Use batch processing systems for scheduling background data processing jobs like quality analysis, production reporting, and maintenance analytics during periods of lower resource demand.

  • Deploy high performance computing (HPC) solutions and run resource-intensive engineering workloads such as computational fluid dynamics (CFD) and computer aided engineering (CAE) simulations during off-peak production hours to optimize resource utilization.

Implementation steps

  1. Assessment and classification:

    • Conduct workload assessment of manufacturing applications using AWS Well-Architected Tool

    • Document peak usage patterns using Amazon CloudWatch

    • Classify applications into real time and batch processing categories

  2. Demand pattern mapping:

    • Create demand heat maps using Amazon CloudWatch metrics

    • Identify off-peak windows for non-time-critical workloads

  3. Resource optimization configuration:

    • Configure AWS Auto Scaling policies with appropriate thresholds

    • Implement scaling plans that align with production schedules

    • Define resource constraints to help prevent over-provisioning

  4. Workload scheduling implementation:

    • Create AWS Batch job configurations for non-critical processing tasks

    • Configure AWS ParallelCluster for engineering simulations during off-hours

    • Implement prioritization logic for computing resources

  5. Monitoring and continuous improvement:

    • Deploy CloudWatch dashboards to track resource utilization efficiency

    • Establish KPIs using AWS Cost and Usage Reports

    • Create sustainability improvement reporting and quarterly review process

Key AWS services

  • AWS Auto Scaling

  • Amazon EC2

  • AWS Batch

  • AWS ParallelCluster

  • Amazon CloudWatch

  • AWS Cost and Usage Reports

Resources