Answer: On-Demand Instances

## Explanation **On-Demand Instances** are the most cost-effective choice for this scenario because: - **Short-term workloads**: On-Demand Instances are billed by the second with no long-term commitments, making them ideal for unpredictable, short-duration workloads - **Irregular patterns**: When usage patterns are inconsistent and cannot be predicted, On-Demand pricing avoids the waste of paying for reserved capacity that goes unused - **Cannot be interrupted**: Unlike Spot Instances (which are not listed here), On-Demand Instances provide guaranteed availability and cannot be interrupted by AWS - **No upfront costs**: There are no upfront payments or long-term contracts required **Why other options are less suitable**: - **Reserved Instances**: Require 1-3 year commitments and are cost-effective only for predictable, steady-state workloads - **Savings Plans**: Also require 1-3 year commitments for consistent usage patterns - **Dedicated Instances**: Provide physical isolation but are more expensive and don't address the irregular workload pattern For workloads that are short-term, irregular, and cannot tolerate interruption, On-Demand Instances provide the optimal balance of flexibility and cost-effectiveness.

Author: Ritesh Yadav