MLS-C01 Question 14
Single answerYou are building a machine learning pipeline that processes large volumes of unstructured image data for training a deep learning model. The data is ingested from multiple sources and needs to be accessed concurrently by multiple training jobs running on Amazon SageMaker. The solution must provide high throughput and low latency for access to the data. Which storage option is the most appropriate for this use case?
- A
Amazon S3
- B
Amazon Elastic File System (Amazon EFS)
- C
Amazon Elastic Block Store (Amazon EBS)
- D
Amazon RDS
Show answer and explanation
Correct answer: B
Explanation
The best storage option for this use case is Amazon Elastic File System (Amazon EFS) because it is designed for low-latency, concurrent access by multiple compute instances, which is necessary for training deep learning models using large image datasets. Amazon S3 is more suitable for object storage and archival, Amazon EBS is limited to a single instance, and Amazon RDS is not suitable for unstructured data like images.
- A. Incorrect.
Amazon S3 is a highly scalable object storage service, but it is not optimized for low-latency, concurrent access by multiple compute instances. It is a better fit for storing large datasets for archival or eventual processing.
- B. Correct.
Amazon Elastic File System (Amazon EFS) is a fully managed, scalable file storage solution that supports concurrent access by multiple compute nodes with low latency. It is ideal for use cases like machine learning training that require shared access to large datasets.
- C. Incorrect.
Amazon Elastic Block Store (Amazon EBS) provides block storage for a single Amazon EC2 instance. While EBS offers high performance, it is not suitable for concurrent access by multiple training jobs.
- D. Incorrect.
Amazon RDS is a managed relational database service. It is not designed to store unstructured data such as images, and it does not support low-latency, concurrent access for machine learning training jobs.