View a markdown version of this page

LSSUS01-BP02 Use energy efficient hardware and services - Life Sciences Lens
Implementation guidanceResources

LSSUS01-BP02 Use energy efficient hardware and services

Select energy-efficient hardware architectures and managed services that optimize power consumption while maintaining research computing performance. Prioritize modern processor architectures and cloud services that incorporate built-in sustainability optimizations to reduce the environmental impact of computational workloads. Implement hardware and service selection strategies that balance performance requirements with energy efficiency goals.

Desired outcome: Achieve significant energy consumption reduction in research computing operations while maintaining or improving computational performance. Implement hardware and service architectures that provide optimal performance-per-watt ratios for life sciences workloads.

Common anti-patterns:

  • You default to traditional x86 architectures without evaluating energy-efficient alternatives.

  • You don't consider managed services that provide built-in sustainability optimizations.

  • You don't evaluate the total cost of ownership including energy consumption.

  • You run custom infrastructure instead of using optimized managed services.

  • You don't monitor and measure energy efficiency metrics across your compute infrastructure.

Benefits of establishing this best practice:

  • Reduce energy consumption by up to 60% through modern processor architectures like AWS Graviton.

  • Lower operational costs through improved performance-per-watt ratios.

  • Improve research productivity with automatically optimized resource allocation.

  • Support organizational sustainability goals and regulatory adherence requirements.

  • Benefit from continuous service improvements and optimizations provided by managed services.

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

Implementation guidance

Energy-efficient hardware selection is crucial for life sciences organizations seeking to reduce their environmental impact while maintaining research computing capabilities. Modern processor architectures like AWS Graviton processors offer significant energy efficiency improvements over traditional x86 architectures, particularly for compute-intensive workloads common in genomics, proteomics, and molecular modeling. These efficiency gains compound over time, making hardware selection a critical sustainability decision.

Managed services provide additional sustainability benefits by incorporating built-in optimizations that individual organizations would struggle to implement independently. Services like AWS HealthOmics, AWS HealthLake, and AWS Batch continuously optimize resource utilization, automatically scale based on demand, and use the latest energy-efficient infrastructure improvements. This approach allows research teams to focus on scientific outcomes while benefiting from ongoing sustainability improvements.

Implementation steps

  1. Evaluate and migrate to energy-efficient processor architectures:

    • Assess workload compatibility with AWS Graviton processors for ARM-based computing.

    • Use Amazon EC2 M6g, C6g, and R6g instances for general-purpose, compute-optimized, and memory-optimized workloads.

    • Benchmark performance and energy consumption for representative research workloads.

    • Consider AWS Graviton3 processors for the latest efficiency improvements.

  2. Use managed services with built-in sustainability optimizations:

    • Migrate genomics workflows to AWS HealthOmics for automated resource optimization.

    • Use AWS HealthLake for healthcare data processing with built-in efficiency features.

    • Implement AWS Batch for containerized workloads with automatic scaling and optimization.

    • Consider Amazon SageMaker AI for machine learning workloads with energy-efficient training.

  3. Optimize service configurations for energy efficiency:

    • Enable AWS Auto Scaling to automatically adjust capacity based on demand.

    • Configure AWS Lambda for event-driven processing to minimize idle resource consumption.

    • Use Amazon ECS with AWS Fargate for serverless container execution.

    • Implement AWS Step Functions for efficient workflow orchestration.

  4. Monitor and measure energy efficiency improvements:

    • Track performance-per-watt metrics using Amazon CloudWatch custom metrics.

    • Monitor cost savings and energy reduction.

    • Implement AWS Config rules to adhere to energy efficiency policies.

  5. Establish continuous optimization processes:

    • Regularly review AWS service updates for new energy efficiency features.

    • Conduct periodic assessments of hardware and service selection decisions.

    • Set up automated alerts for suboptimal resource utilization patterns.

    • Create feedback loops to incorporate energy efficiency learnings into future architecture decisions.

Resources

Related best practices:

Related documents:

Related videos:

Related examples: