MLSUS02-BP02 Select sustainable Regions
Choose the Regions where you implement your workloads based on both your business requirements and sustainability goals.
Desired outcome: You select AWS Regions that align with your organizational sustainability objectives while meeting your business requirements. By choosing Regions with renewable energy sources and lower carbon intensity, you reduce the environmental impact of your machine learning workloads while maintaining optimal performance for your business needs.
Common anti-patterns:
-
Selecting Regions based solely on proximity without considering environmental impact.
-
Ignoring renewable energy availability when deploying machine learning workloads.
-
Deploying workloads across multiple Regions without considering their carbon footprints.
Benefits of establishing this best practice:
-
Alignment with organizational sustainability goals and ESG initiatives.
-
Enhanced reputation as an environmentally responsible organization.
-
Potential cost savings through efficient Region selection.
Level of risk exposed if this best practice is not established: Medium
Implementation guidance
When deploying your machine learning workloads, Region selection plays a crucial role in meeting both your operational requirements and sustainability goals. While factors such as latency, data residency, and service availability remain important, incorporating sustainability considerations into your Region selection process can minimize your environmental impact. AWS is continuously expanding its renewable energy projects globally, making it increasingly possible to host your workloads in Regions powered by sustainable energy sources.
The cloud offers significant sustainability advantages compared to on-premises deployments due to higher utilization rates, more energy-efficient infrastructure, and AWS' commitment to renewable energy. By selecting Regions thoughtfully, you can further enhance these sustainability benefits while still meeting your business needs.
Implementation steps
-
Understand your business requirements first. Identify the non-negotiable requirements for your workload, including data sovereignty regulations, compliance-aligned needs, latency requirements, and service availability in specific Regions. Create a shortlist of Regions that meet these baseline requirements.
-
Research AWS renewable energy projects. Use the Amazon Around the Globe
resource to identify Regions that are near Amazon renewable energy projects. AWS achieved powering its operations with 100% renewable energy in 2023, seven years ahead of their original 2030 commitment. -
Consider the grid's carbon intensity. Look for Regions where the electrical grid has lower published carbon intensity. This information may be available through regional utility reports or sustainability documentation. Lower carbon intensity means reduced emissions even for non-renewable energy sources.
-
Evaluate the trade-offs. When selecting Regions, consider potential trade-offs between sustainability goals and business requirements such as latency or availability. In some cases, minor performance trade-offs may be acceptable to achieve significant sustainability improvements.
-
Monitor sustainability metrics. After deployment, track relevant sustainability metrics to verify that your Region selection is delivering the expected environmental benefits. Consider implementing dashboards with key performance indicators (KPIs) for sustainability tracking.
-
Review and adjust periodically. As AWS adds more renewable energy projects and as your business requirements evolve, periodically reassess your Region selections to continually align with your sustainability goals.
Resources
Related documents:
