AWS Certified Machine Learning – SpecialtyMachine Learning Implementation and OperationsEasy
A data science team is deploying a new machine learning model for real-time inference on Amazon SageMaker. They need to ensure that if an Availability Zone (AZ) becomes unhealthy, inference traffic is automatically routed to healthy AZs with minimal downtime. Which SageMaker deployment configuration best addresses this requirement?
- ADeploying the model using SageMaker Batch Transform for high availability.
- BDeploying the model to a single SageMaker endpoint with auto-scaling enabled.
- CDeploying the model to multiple SageMaker endpoints in different regions.
- DDeploying the model to a multi-AZ SageMaker endpoint configuration.
Show answer & explanationAnswer & explanation
Correct answer: D. Deploying the model to a multi-AZ SageMaker endpoint configuration.
A multi-AZ SageMaker endpoint configuration automatically distributes and manages instances across multiple Availability Zones. This ensures that if one AZ experiences issues, traffic is seamlessly rerouted to healthy AZs, providing high availability and fault tolerance for real-time inference.
Why the other options are wrong
- A. SageMaker Batch Transform is for offline, asynchronous inference and is not suitable for real-time, low-latency requirements.
- B. Auto-scaling helps with traffic spikes but does not inherently provide cross-AZ fault tolerance for the endpoint itself.
- C. Deploying to multiple regions provides disaster recovery for regional outages, but not automatic failover for AZ issues within a single region, and adds latency.
SageMaker Multi-AZ Endpoint
A SageMaker endpoint configuration that deploys model instances across multiple AWS Availability Zones to provide high availability and fault tolerance for real-time inference.
- Ensures continuous inference even if one AZ fails.
- Automatically distributes traffic and reroutes from unhealthy AZs.
- Critical for production workloads requiring high uptime.
Memory trick: Many Zones, Always On, SageMaker's Strong.