AWS Certified Solutions Architect – Professional (SAP-C02) — Question 402

A company is deploying a distributed in-memory database on a fleet of Amazon EC2 instances. The fleet consists of a primary node and eight worker nodes. The primary node is responsible for monitoring cluster health, accepting user requests, distributing user requests to worker nodes, and sending an aggregate response back to a client. Worker nodes communicate with each other to replicate data partitions.

The company requires the lowest possible networking latency to achieve maximum performance.

Which solution will meet these requirements?

Answer options

• A. Launch memory optimized EC2 instances in a partition placement group.
• B. Launch compute optimized EC2 instances in a partition placement group.
• C. Launch memory optimized EC2 instances in a cluster placement group.
• D. Launch compute optimized EC2 instances in a spread placement group.

Correct answer: C

Explanation

For a distributed in-memory database, memory optimized EC2 instances are required to handle the large memory footprint and throughput demands. To achieve the absolute lowest network latency and highest throughput between these instances, they must be deployed in a cluster placement group, which locates instances physically close together on the same underlying hardware. Partition and spread placement groups are designed to reduce correlated hardware failures by spreading instances across different racks, which increases latency.