MLA-C01 Question 39
Single answerYou are building a machine learning pipeline that processes large amounts of unstructured data, such as raw images and videos. The data must be stored cost-effectively and accessed frequently during the training phase but less frequently during the evaluation phase. Which AWS storage solution is the most suitable for this scenario, considering both cost and performance tradeoffs?
- A
Amazon S3 Standard
- B
Amazon S3 Glacier Deep Archive
- C
Amazon FSx for Lustre
- D
Amazon EBS General Purpose SSD (gp3)
Show answer and explanation
Correct answer: A
Explanation
Amazon S3 Standard offers the right balance of cost-effectiveness, scalability, and frequent access performance, making it the best choice for storing large amounts of unstructured data during the training and evaluation phases of a machine learning pipeline. Other options like Amazon FSx for Lustre and Amazon EBS provide high performance but are less cost-efficient for large-scale storage, while Amazon S3 Glacier Deep Archive is designed for infrequent access and archival use cases.
- A. Correct.
Amazon S3 Standard is a highly cost-effective and scalable object storage solution designed for frequent access. It is ideal for storing unstructured data like images and videos during the training phase.
- B. Incorrect.
Amazon S3 Glacier Deep Archive is optimized for long-term archival storage and infrequent access. It is not suitable for frequent access, especially in a machine learning pipeline.
- C. Incorrect.
Amazon FSx for Lustre is a high-performance file system designed for high-throughput, low-latency workloads. While it is suitable for training, it is less cost-effective for storing large amounts of raw data compared to Amazon S3.
- D. Incorrect.
Amazon EBS General Purpose SSD (gp3) is a block storage option designed for low-latency, high-performance applications. It is not cost-effective for storing large-scale unstructured data like images and videos.