Q1: What is Amazon Elastic Inference?
A1: Amazon Elastic Inference is a service provided by AWS that allows customers to attach low-cost GPU-powered acceleration to Amazon EC2, Amazon SageMaker instances, or Amazon Elastic Container Service (ECS) tasks. It helps reduce the cost of running deep learning inference by up to 75% compared to standalone GPU-based instances.
Q2: Why is Amazon encouraging customers to move workloads from Amazon Elastic Inference (EI) to newer hardware acceleration options such as AWS Inferentia?
A2: Amazon is encouraging customers to move workloads from Amazon Elastic Inference to newer hardware acceleration options like AWS Inferentia because these options provide better performance at a much better price for inference workloads. AWS Inferentia is designed to provide high-performance inference in the cloud and drive down the total cost of inference.
Q3: Which AWS services are impacted by the move to stop onboarding new customers to Amazon Elastic Inference (EI)?
A3: The move to stop onboarding new customers to Amazon Elastic Inference impacts Amazon EC2, Amazon SageMaker instances, and Amazon Elastic Container Service (ECS) tasks. This applies to both endpoints and notebook kernels using Amazon Elastic Inference accelerators.
Q4: Will I be able to create a new Amazon Elastic Inference (EI) accelerator after April 15, 2023?
A4: If you are a new customer and have not used Amazon Elastic Inference in the past 30 days, you will not be able to create a new Amazon Elastic Inference instance in your AWS account after April 15, 2023. However, if you have used an Amazon Elastic Inference accelerator at least once in the past 30 days, you can attach a new accelerator to your instance.
Q5: How do I evaluate alternative instance options for my current Amazon SageMaker Inference Endpoints?
A5: You can use Amazon SageMaker Inference Recommender to help identify cost-effective deployments for migrating existing workloads from Amazon Elastic Inference to an appropriate ML instance supported by SageMaker. This will help you choose the best alternative instance options for your current endpoints.