A company wants to run an in-memory database for a latency-sensitive application that runs on Amazon ECz2 instances. The application processes more than 100,000 transactions each minute and requires high network throughput. A solutions architect needs to provide a cost-effective network design that minimizes data transfer charges. Which solution meets these requirements?

1. Launch all EC2 instances in the same Availability Zone within the same AWS Region. Specify a placement group with cluster strategy when launching EC2 instances.
2. Launch all EC2 instances in different Availability Zones within the same AWS Region. Specify a placement group with partition strategy when launching EC2 instances.
3. Deploy an Auto Scaling group to launch EC2 instances in different Availability Zones based on a network utilization target.
4. Deploy an Auto Scaling group with a step scaling policy to launch EC2 instances in different Availability Zones.