AWS Certified Solutions Architect – Professional — Question 286

A solutions architect needs to migrate 50 TB of NFS data to Amazon S3. The files are on several NFS file servers on corporate network. These are dense file systems containing tens of millions of small files. The system operators have configured the file interface on an AWS Snowball Edge device and are using a shell script to copy data.
Developers report that copying the data to the Snowball Edge device is very slow. The solutions architect suspects this may be related to the overhead of encrypting all the small files and transporting them over the network.
Which changes can be made to speed up the data transfer?

Answer options

• A. Cluster two Snowball Edge devices together to increase the throughput of the devices.
• B. Change the solution to use the S3 Adapter instead of the file interface on the Snowball Edge device.
• C. Increase the number of parallel copy jobs to increase the throughput of the Snowball Edge device.
• D. Connect directly to the USB interface on the Snowball Edge device and copy the files locally.

Correct answer: C

Explanation

Transferring millions of small files sequentially introduces significant overhead due to metadata operations and network latency. Increasing the number of parallel copy jobs allows the AWS Snowball Edge to process multiple files simultaneously, which maximizes the device's throughput capabilities. Other options, such as using the USB interface or clustering, do not resolve the software-level bottleneck of single-threaded file transfers.