SOA-C02 Question 339
Select 3A company is running a high-performance computing (HPC) workload on Amazon EC2 and requires low latency, high throughput, and optimized networking between instances. Which configuration steps should the SysOps Administrator take to enable these enhanced EC2 capabilities?
- A
Launch instances in a cluster placement group.
- B
Attach an Elastic Network Adapter (ENA) to the instances.
- C
Use EBS-backed volumes for maximum IOPS performance.
- D
Select an instance type that supports enhanced networking.
- E
Enable EC2 Hibernate for faster instance recovery.
Show answer and explanation
Correct answers: A, B, D
Explanation
To optimize EC2 instances for HPC workloads, the SysOps Administrator must configure enhanced networking and low-latency placement. Using a cluster placement group ensures physical proximity for lower latency. ENA provides enhanced networking capabilities such as higher bandwidth and lower latency. Additionally, the instance type must support these features for them to work. These configurations together fulfill the requirements of HPC workloads.
- A. Correct.
Launching instances in a cluster placement group ensures low latency and high throughput by placing instances physically close to each other in the same Availability Zone. This is a key requirement for HPC workloads.
- B. Correct.
Attaching an Elastic Network Adapter (ENA) enables enhanced networking, which provides higher bandwidth and lower latency, critical for HPC workloads.
- C. Incorrect.
Using EBS-backed volumes is unrelated to networking and does not optimize network throughput or latency. Instance store or other storage configurations may be more relevant for HPC but aren't tied to networking improvements.
- D. Correct.
Selecting an instance type that supports enhanced networking is essential for enabling features like ENA. Enhanced networking is crucial for achieving high performance in HPC scenarios.
- E. Incorrect.
EC2 Hibernate allows instances to preserve their state, but it is unrelated to networking or placement group requirements for HPC workloads.