Which architecture can the company use to implement the required A/B testing setup?
Create two new SageMaker endpoints for the TensorFlow and PyTorch models in addition to the existing SageMaker endpoint. Create an Application Load Balancer. Create a target group for each endpoint. Configure listener rules and add weight to the target groups. To send traffic to the TensorFlow model for customers who are from Europe, create an additional listener rule to forward traffic to the TensorFlow target group.
Create two production variants for the TensorFlow and PyTorch models. Create an auto scaling policy and configure the desired A/B weights to direct traffic to each production variant. Update the existing SageMaker endpoint with the auto scaling policy. To send traffic to the TensorFlow model for customers who are from Europe, set the TargetVariant header in the request to point to the variant name of the TensorFlow model.
Create two new SageMaker endpoints for the TensorFlow and PyTorch models in addition to the existing SageMaker endpoint. Create a Network Load Balancer. Create a target group for each endpoint. Configure listener rules and add weight to the target groups. To send traffic to the TensorFlow model for customers who are from Europe, create an additional listener rule to forward traffic to the TensorFlow target group.
Create two production variants for the TensorFlow and PyTorch models. Specify the weight for each production variant in the SageMaker endpoint configuration. Update the existing SageMaker endpoint with the new configuration. To send traffic to the TensorFlow model for customers who are from Europe, set the TargetVariant header in the request to point to the variant name of the TensorFlow model.
Explanations:
The use of an Application Load Balancer (ALB) is not suitable for SageMaker model deployment. SageMaker endpoints do not integrate directly with ALBs for model traffic management, and the need for routing traffic based on geographic location adds complexity that is more suited for SageMaker’s built-in functionality.
While creating production variants and setting traffic weights is valid, the TargetVariant header is not the appropriate method to route traffic based on customer location (Europe). SageMaker doesn’t support geographic-based routing directly via this method.
Similar to Option A, using a Network Load Balancer (NLB) with SageMaker endpoints is not the ideal approach for A/B testing in this case. Furthermore, geographic traffic routing based on customer location would be more complicated and less efficient with NLB in this context.
This option correctly leverages SageMaker’s native ability to create production variants with traffic weighting, directly modifying the existing endpoint to route traffic based on weights. Additionally, SageMaker allows for configuration to handle geographic routing, which makes this setup ideal for the company’s A/B testing needs.
I infer that the answer is:
Create two production variants for the TensorFlow and PyTorch models. Specify the weight for each production variant in the SageMaker endpoint configuration. Update the existing SageMaker endpoint with the new configuration. To send traffic to the TensorFlow model for customers who are from Europe, set the TargetVariant header in the request to point to the variant name of the TensorFlow model.