SAA-C03 Practice Question 46: FSx Lustre HPC Throughput
AWS Certified Solutions Architect – Associate (SAA-C03) · Test 5 · question 6 of 10
A genomics pipeline runs on a fleet of EC2 instances that must read a multi-terabyte dataset held in Amazon S3. The compute nodes need a POSIX file system with sub-millisecond latency and hundreds of gigabytes per second of aggregate throughput for the duration of each run, after which the results are written back to S3 and the file system is discarded.
Which storage service fits these requirements?
Show answer and explanation
Answer:
FSx for Lustre is the high-performance parallel file system AWS offers for exactly this workload: sub-millisecond latencies and throughput scaling into hundreds of gigabytes per second. It links directly to an S3 bucket, presenting objects as files and loading data lazily as it is read, and results can be exported back to S3. A scratch deployment matches the short-lived pattern described, since the file system is meant to be created for a run and deleted afterwards.
Option B provides a shared POSIX file system but not this performance profile. EFS is built for general-purpose shared access, and Max I/O raises throughput at the cost of higher per-operation latency, which is the opposite of the sub-millisecond requirement.
Option C is block storage attached to one instance at a time, so it is not a shared file system for the fleet, and synchronising terabytes to every node duplicates the data and the effort.
Option D serves Windows workloads over SMB. This pipeline needs POSIX semantics and HPC throughput, which is what the Lustre offering rather than the Windows one provides.
Reference: What is Amazon FSx for Lustre?