AWS Certified Solutions Architect Professional SAP-C02 Practice Question
A risk-analytics team launches a short-lived, 10,000-node Monte Carlo simulation on EC2 Spot instances in a single AWS Region. During the 30-minute run each node concurrently creates, reads, and deletes millions of 2 MB files. The application requires POSIX semantics, sub-millisecond latency, and an aggregate throughput that can scale to hundreds of GB/s. After the simulation finishes, results are copied to Amazon S3 and the scratch storage can be discarded. To maximize performance and pay only for storage used during the job, which AWS storage option should the solutions architect choose?
Store intermediate data in an Amazon S3 Express One Zone directory bucket that is co-located with the compute Availability Zone.
Create an Amazon FSx for Lustre file system using the Scratch 2 deployment type with the SSD storage class, mount it on all compute nodes, and delete the file system when the simulation completes.
Deploy an Amazon FSx for Lustre Persistent 2 file system with 1,000 MB/s-per-TiB throughput and SSD storage class, keeping it online between runs.
Provision an Amazon EFS Regional file system in Elastic throughput mode and General Purpose performance mode for the cluster.
Amazon FSx for Lustre scratch file systems are purpose-built for temporary, highly parallel HPC workloads. The Scratch 2 deployment type delivers 200 MB/s of baseline throughput per TiB and can burst to six times that value, providing hundreds of GB/s (and even TB/s) of aggregate throughput while maintaining sub-millisecond latency. Because data is not replicated and you are billed only for the storage provisioned while the file system exists, deleting the file system after the job minimizes cost.
Amazon EFS with Elastic throughput tops out at tens of GiB/s, far below the required throughput. S3 Express One Zone offers high request rates but lacks POSIX file-system semantics. An FSx for Lustre Persistent 2 file system would meet the performance goal but adds redundant storage, per-TiB-month charges, and throughput capacity fees that are unnecessary for disposable scratch data, making it less cost-optimal.
Ask Bash
Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.
Why is Amazon FSx for Lustre Scratch 2 more suitable than Persistent 2 for this workload?
Open an interactive chat with Bash
What does 'POSIX semantics' mean in the context of AWS storage services?
Open an interactive chat with Bash
Why is Amazon EFS not suitable for this type of workload?
Open an interactive chat with Bash
AWS Certified Solutions Architect Professional SAP-C02
Design for New Solutions
Your Score:
Report Issue
Bash, the Crucial Exams Chat Bot
AI Bot
Loading...
Loading...
Loading...
IT & Cybersecurity Package Join Premium for Full Access