Exam AWS Certified Machine Learning - Specialty topic 1 question 206 discussion

CE Option A: Launching multiple medium-sized instances in a distributed SageMaker Processing job and using the prebuilt Docker images for Apache Spark to query and plot the relevant data is a possible solution, but it may not be the most cost-effective solution as it requires spinning up multiple instances. Option B: Launching multiple medium-sized notebook instances with a PySpark kernel in distributed mode is another solution, but it may not be the most secure solution as the data would be stored on the instances and not in a centralized data repository. Option D: Using AWS Secrets Manager to store the Amazon Redshift credentials and launching a SageMaker extra-large notebook instance is a solution, but the block storage requirement that is slightly larger than 10 TB could be costly and may not be necessary.

C and E. No secure control is in option A.

Option A can do it as well but could be expensive and not as easy as option E

As soon as I see, Download andpython client, I am worried about speed and efficiency. So I would say A and E

A. NO - SageMaker Processing job is a self-contained feature using the sagemaker.processing API; it does not rely on invoking Spark directly B. NO - you want to identify the relevant slice of data without having to download everything first C. YES - minimize data movement D. NO - you want to identify the relevant slice of data without having to download everything first E. YES - built-in tool specifically designed for that use case

E for sure but was a bit confused on A or C but based on the link would go for C https://aws.amazon.com/blogs/big-data/using-the-amazon-redshift-data-api-to-interact-from-an-amazon-sagemaker-jupyter-notebook/

e is obvious choice: c https://aws.amazon.com/blogs/big-data/using-the-amazon-redshift-data-api-to-interact-from-an-amazon-sagemaker-jupyter-notebook/

C & E seems right - https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler.html