A company needs to run an in‑memory database for a latency‑sensitive application on Amazon EC2. The app processes over 100,000 transactions per minute and requires high network throughput. A solutions architect must design a cost‑effective network layout that minimizes data transfer charges. Which solution meets these requirements?

The correct solution leverages a cluster placement group within a single Availability Zone (AZ). Cluster placement groups pack instances closely together, providing ultra-low latency and high network throughput, which is critical for a latency-sensitive, high-transaction in-memory database. Keeping all instances in the same AZ minimizes data transfer costs, as inter-AZ data transfer incurs charges, while intra-AZ transfer between instances in a placement group is free. Launching instances in different AZs (partition strategy or Auto Scaling Group) would increase latency and data transfer costs due to inter-AZ communication, which is detrimental to a latency-sensitive application. Auto Scaling Groups are beneficial for scalability but don't directly address the ultra-low latency and high throughput networking requirements as effectively as a cluster placement group for this specific workload.