MLA-C01 Question 310
Select 3A company is developing a machine learning model and needs to deploy it as a REST API for inference. They want to use a container-based solution to ensure scalability, ease of deployment, and integration with other AWS services. Which AWS services should the company consider using to achieve their goal?
- A
Amazon Elastic Kubernetes Service (EKS)
- B
AWS Lambda
- C
Amazon Elastic Container Service (ECS)
- D
AWS Fargate
- E
Amazon SageMaker Hosting Services
Show answer and explanation
Correct answers: A, C, D
Explanation
To deploy a machine learning model as a REST API using a container-based solution, the company can use Amazon EKS, Amazon ECS, or AWS Fargate. These services are designed to manage containerized applications and integrate seamlessly with other AWS services. While AWS Lambda supports container images, it is not ideal for long-running inference tasks. Amazon SageMaker Hosting Services is specific to ML model hosting and lacks the general-purpose container orchestration capabilities offered by EKS, ECS, or Fargate.
- A. Correct.
Amazon Elastic Kubernetes Service (EKS) is a fully managed Kubernetes service that allows you to deploy and scale containerized applications, including machine learning inference services. It is well-suited for scenarios requiring advanced orchestration and scalability.
- B. Incorrect.
AWS Lambda is a serverless compute service, but it is not specifically designed for running containerized applications in a scalable manner. While it supports container images, it is not ideal for long-running machine learning inference workloads.
- C. Correct.
Amazon Elastic Container Service (ECS) is a fully managed container orchestration service that simplifies the deployment and scaling of containerized applications. It integrates well with other AWS services and is a suitable choice for deploying machine learning inference APIs.
- D. Correct.
AWS Fargate is a serverless compute engine that works with ECS and EKS to run containers without managing the underlying servers. It is an excellent choice for running containerized applications, including machine learning inference APIs, without worrying about infrastructure management.
- E. Incorrect.
Amazon SageMaker Hosting Services is specifically designed for deploying and managing machine learning models, but it is not a general-purpose container orchestration service. While it can serve models, it does not provide the flexibility of managing your own containerized applications for REST APIs.