AWS HPC Networking and Storage Solutions
Short answer
BD Use Amazon FSx for Lustre for high-throughput scratch storage and attach Elastic Fabric Adapters (EFA) to EC2 instances for low-latency inter-node communication.
Question
Key Takeaways
The question tests the ability to distinguish between general-purpose AWS services and specialized HPC infrastructure components, with the common trap being the selection of services like CloudFront or Global Accelerator which do not address intra-cluster communication needs.
This question addresses the design of tightly coupled High Performance Computing (HPC) environments by identifying the optimal AWS services for low-latency networking and high-throughput storage. It establishes that Amazon FSx for Lustre and Elastic Fabric Adapters (EFA) are the specific solutions required to meet these performance requirements.
Common mistake: Learners often incorrectly select options related to content delivery or application deployment, such as CloudFront or Elastic Beanstalk, failing to recognize that 'tightly coupled' workloads require direct, low-latency inter-node communication rather than external access optimization.
Expert Analysis
Why the Answer Is Correct
Tightly coupled HPC workloads rely on frequent, low-latency communication between nodes, typically using Message Passing Interface (MPI). Option D is correct because Elastic Fabric Adapters (EFA) are specifically designed to enhance network performance for HPC applications by providing low-latency and high-throughput networking capabilities directly integrated with EC2 instances. Option B is correct because Amazon FSx for Lustre is a file system optimized for high-performance computing, offering the massive throughput and low latency required for shared storage in these environments.
Why the Other Options Are Wrong
Option A is incorrect because AWS Global Accelerator improves availability and performance for internet-facing applications by routing traffic to endpoints, not for internal cluster communication. Option C is incorrect because Amazon CloudFront is a Content Delivery Network (CDN) used for caching content at edge locations for end-users, not for inter-instance data exchange. Option E is incorrect because Elastic Beanstalk is an Application Platform as a Service (PaaS) for deploying applications, lacking the granular control and specialized networking features needed for raw HPC compute clusters.
Community Comment Notes
Community consensus strongly supports BD, with users noting that EFA provides improved ENA for Linux-based HPC workloads. One commenter highlighted that EFA leverages MPI standards to bypass the Linux OS for reliable transport. Another user clarified that Global Accelerator focuses on geographically distributed users, making it unsuitable for intra-cluster communication.
Official Reference
- https://docs.aws.amazon.com/efs/latest/ug/what-is-efs.html
- https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/efa.html
These links open the official vendor documentation in a new tab.
Exam Strategy
When designing HPC architectures, always look for keywords like 'tightly coupled,' 'low latency,' and 'high performance.' Associate these with EFA for networking and FSx for Lustre for storage, as these are the standard AWS offerings for this specific workload type.
Frequently Asked Questions
Why is Global Accelerator wrong for HPC?
Can I use EBS instead of FSx for Lustre?
Community Votes
100% of anonymous learners picked answer BD. Votes are pick records left by other test-takers — they are not the verified answer.
Discussion 4 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 4
-
asdfcdsxdfc 👍 4Elastic Fabric Adapter (EFA) • Improved ENA for HPC, only works for Linux • Great for inter-node communications, tightly coupled workloads • Leverages Message Passing Interface (MPI) standard • Bypasses the underlying Linux OS to provide low-latency, reliable transport
-
LeonSauveterre 👍 1A - Global Accelerator is not designed for tightly coupled HPC environments. It focuses on improving performance for geographically distributed users, not for low-latency intra-cluster communication. B - FSx for Lustre is specifically designed for HPC workloads and integrates well with Amazon S3 for high-throughput and low-latency access to data. C - CloudFront does not address the low-latency, high-bandwidth networking or storage requirements of tightly coupled HPC environments. D - EFA is a network interface that provides low-latency, high-throughput connectivity for tightly coupled HPC workloads, such as those using MPI (Message Passing Interface). E - Elastic Beanstalk is used for deploying and managing web applications and services, which has nothing to do with HPC.
-
Andy_09 👍 7Options BD
-
seetpt 👍 3BD looks right
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.
Related Analysis
Practice All SAA-C03 Questions
Access 120 questions with complete answers and detailed explanations.
View Full SAA-C03 Practice Test →