MLA-C01 Question 288
Select 3You are tasked with setting up the infrastructure for a machine learning system that processes real-time streaming data and performs inference using a pre-trained model. The requirements specify the following:
- The system must scale automatically based on streaming data volume.
- The pre-trained model should be deployed as a fully managed service.
- The solution must minimize operational overhead.
Which combination of AWS services should you use to create and script the infrastructure based on these requirements?
- A
Amazon Kinesis Data Streams for real-time data ingestion and AWS Lambda for preprocessing the data
- B
Amazon S3 for real-time data ingestion and Amazon EC2 for hosting the pre-trained model
- C
Amazon SageMaker Endpoint for deploying the pre-trained model and making real-time predictions
- D
Amazon EC2 Auto Scaling for scaling compute resources and AWS Batch for preprocessing the data
- E
AWS CloudFormation to script and automate the deployment of the entire infrastructure
Show answer and explanation
Correct answers: A, C, E
Explanation
The correct combination of services leverages Amazon Kinesis Data Streams for real-time data ingestion, AWS Lambda for preprocessing, Amazon SageMaker Endpoint for managed model deployment, and AWS CloudFormation for infrastructure scripting and automation. This setup satisfies all the requirements: real-time data processing, minimal operational overhead, and automated scaling.
- A. Correct.
Amazon Kinesis Data Streams efficiently handles real-time data ingestion, and AWS Lambda provides serverless preprocessing, meeting the requirement for automatic scaling and minimizing operational overhead.
- B. Incorrect.
Amazon S3 is not suitable for real-time data ingestion, and using Amazon EC2 for hosting the model would increase operational overhead compared to a managed service like SageMaker.
- C. Correct.
Amazon SageMaker Endpoint allows you to deploy the pre-trained model as a fully managed service, meeting the requirement for minimizing operational overhead and simplifying real-time inference.
- D. Incorrect.
While Amazon EC2 Auto Scaling can scale compute resources, AWS Batch is not designed for real-time data preprocessing, making this combination unsuitable for the given requirements.
- E. Correct.
AWS CloudFormation enables you to script and automate the deployment of the entire infrastructure, ensuring consistency and reducing manual effort, which aligns with the requirements.