MLA-C01 Question 272
Single answerYou are deploying a machine learning model using Amazon SageMaker and need to handle predictions for large batches of data submitted periodically by your users. Which type of endpoint should you choose and why?
- A
Serverless endpoint
- B
Real-time endpoint
- C
Asynchronous endpoint
- D
Batch transform
Show answer and explanation
Correct answer: D
Explanation
Batch transform is the optimal choice for processing large datasets in periodic batches. It is designed for offline inference, where predictions do not need to be served in real-time. This solution is scalable, efficient, and avoids the overhead of maintaining an endpoint continuously, unlike serverless or real-time endpoints, which are more suited for real-time or low-latency requirements.
- A. Incorrect.
Serverless endpoints are designed for applications with low and intermittent traffic, providing real-time responses without managing infrastructure. They are not suitable for large batch processing.
- B. Incorrect.
Real-time endpoints are ideal for low-latency predictions with immediate response requirements. However, they are not efficient for handling large batches of data submitted periodically.
- C. Incorrect.
Asynchronous endpoints can handle large payloads and are suitable for scenarios where immediate responses are not required. However, they are better suited for individual predictions rather than processing large batches at once.
- D. Correct.
Batch transform is specifically designed for processing large datasets in batches. It is optimized for periodic, large-scale inference jobs, making it the best choice for this scenario.