Question #779
A company is deploying a distributed database that requires consistent high-throughput and low-latency communication between five Amazon EC2 instances within the same AWS Region. The solution does not need to account for fault tolerance.
Which architecture will satisfy these requirements?
Deploy five EC2 instances within a cluster placement group, using an instance type that supports enhanced networking capabilities.
Deploy five EC2 instances in an Auto Scaling group across a single Availability Zone, attaching an additional elastic network interface to each instance.
Deploy five EC2 instances within a partition placement group, selecting an instance type that supports enhanced networking capabilities.
Deploy five EC2 instances across a spread placement group, attaching an additional elastic network interface to each instance.
Explanation
The correct answer is A because:
- Cluster Placement Groups: Designed for applications requiring the lowest latency and highest throughput. By placing instances in the same Availability Zone on adjacent hardware, network performance is maximized.
- Enhanced Networking: Instance types supporting enhanced networking (e.g., SR-IOV) reduce latency, increase packet-per-second (PPS) performance, and improve throughput.
- Fault Tolerance Not Required: Since fault tolerance is not a concern, the trade-off of reduced fault tolerance (due to all instances being in the same AZ/hardware) is acceptable.
Why other options are incorrect:
- B: Auto Scaling groups and additional ENIs do not guarantee low-latency communication. Instances may still be distributed across different hardware.
- C: Partition placement groups prioritize fault tolerance by isolating partitions, which introduces latency between partitions and is unnecessary here.
- D: Spread placement groups maximize fault tolerance by spreading instances across hardware, increasing latency and reducing throughput.
Key Points:
- Use cluster placement groups for low-latency/high-throughput workloads when fault tolerance is not required.
- Enhanced networking optimizes EC2 network performance.
- Avoid spread/partition placement groups if fault tolerance is not a priority.
Answer
The correct answer is: A