A weather forecasting company needs to process hundreds of gigabytes of data with sub-millisecond latency. The company has a high performance…

The scenario describes a High-Performance Computing (HPC) workload that requires sub-millisecond latency, high sustained throughput, and massive parallel access from thousands of compute instances. Amazon FSx for Lustre is the AWS service specifically designed for these demanding HPC requirements. The question also specifies the need for a "highly available" solution. FSx for Lustre offers two deployment types: persistent and scratch. Persistent file systems are designed for longer-term workloads, providing high availability by replicating data and automatically replacing failed file servers within an Availability Zone. This directly meets all the requirements of the scenario.

A. Use Amazon FSx for Lustre scratch file systems: Scratch file systems are for temporary processing and are not highly available. Data is not replicated and is lost if a file server fails.

C. Use Amazon Elastic File System (Amazon EFS) with Bursting Throughput mode: EFS is a general-purpose file system. Its latency is typically in the low single-digit milliseconds, not sub-millisecond. Bursting mode is not suitable for sustained high throughput.

D. Use Amazon Elastic File System (Amazon EFS) with Provisioned Throughput mode: While providing sustained throughput, EFS is not optimized for the extreme low latency and massive parallel I/O demands of HPC workloads like FSx for Lustre is.

---

1. Amazon FSx for Lustre Documentation

"What is Amazon FSx for Lustre?": "Amazon FSx for Lustre is a fully managed service that provides cost-effective

high-performance

scalable storage for compute workloads. ... FSx for Lustre is ideal for high-performance computing (HPC)... with latencies of a few hundred microseconds." This confirms FSx for Lustre is the correct service for the HPC use case and sub-millisecond latency requirement.

2. Amazon FSx for Lustre Documentation

"Deployment options for FSx for Lustre file systems": "Persistent file systems are designed for longer-term storage and workloads. The file servers are highly available

and data is automatically replicated within the same Availability Zone... Scratch file systems are ideal for temporary storage and shorter-term processing of data. Data is not replicated and does not persist if a file server fails." This reference directly supports choosing the "persistent" option to meet the "highly available" requirement.

3. Amazon EFS Documentation

"Amazon EFS performance": "Amazon EFS offers two performance modes: General Purpose and Max I/O... For all file systems

Amazon EFS is designed to provide consistent

low latencies. You can expect low single-digit millisecond latencies for most file operations." This shows EFS latency is higher than the sub-millisecond latency provided by FSx for Lustre.

4. AWS Solutions

"High Performance Computing (HPC) on AWS": This page positions Amazon FSx for Lustre as a key storage component for HPC

stating it provides "a high-performance file system optimized for fast processing of workloads." This reinforces its suitability for the described scenario over a general-purpose file system like EFS.