SOA-C02 Cost and Performance Optimization Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Use Dedicated Hosts to run the application and share resources across multiple accounts.

  Why it's wrong here

  Dedicated Hosts are physically isolated servers for your exclusive use, typically required for server-bound software licenses such as Windows Server with specific licensing agreements. They incur an hourly cost regardless of whether the instance is utilized, and they do not provide any inherent mechanism for sharing resources across multiple AWS accounts or for automatically adjusting capacity to match variable web traffic. Sharing across accounts is better achieved with AWS Resource Access Manager or VPC sharing — not by spending on Dedicated Hosts.

• ✓

  Use a mix of On-Demand and Spot Instances in an Auto Scaling group with a target tracking scaling policy.

  Why this is correct

  A mixed-instance Auto Scaling group with On-Demand and Spot Instances is cost-optimal because Spot Instances are available for up to 90% lower hourly price, and the target tracking scaling policy automatically adjusts the desired capacity based on a selected metric such as average CPU utilization, maintaining a baseline with On-Demand while absorbing traffic spikes with Spot. The ASG's capacity rebalancing feature monitors Spot interruption warnings and proactively launches replacement instances, making this an ideal, resilient and inexpensive solution for a fault-tolerant web tier.

• ✗

  Purchase Reserved Instances for the expected baseline capacity and use On-Demand for spikes.

  Why it's wrong here

  Reserved Instances provide significant discounts over On-Demand but require a 1- or 3-year commitment to specific instance attributes, making them suitable only for steady-state, predictable workloads with constant usage. For a web application with variable traffic, predicting the exact baseline capacity is difficult; over-purchasing RIs wastes spending and under-purchasing defeats the purpose, and they cannot elastically scale up or down in response to load, so they are less flexible than a dynamically-adjusted Auto Scaling group with Spot capacity.

• ✗

  Use only On-Demand EC2 Instances with an Auto Scaling group to handle variable traffic.

  Why it's wrong here

  While an Auto Scaling group with only On-Demand Instances will correctly handle variable traffic by adding and removing instances based on load, it ignores a key cost-optimization opportunity: the application is fault-tolerant and can easily run a substantial portion of its workload on Spot Instances at a fraction of the price. The target tracking policy works equally well with mixed instance types, but selecting only On-Demand means every single instance is procured at full list price, making the architecture unnecessarily expensive. A cost-effective design must use the lowest-priced purchasing options that still meet performance and availability requirements, and On-Demand-only capacity does not maximize the savings available from Spot.

Go deeper

Related to this question