You have EC2 instances that run continuously and expect roughly the same usage for at least the next year. To reduce cost compared to on-demand pricing, which purchasing option is the most appropriate?

Explanation

For workloads with stable long-term usage, Reserved Instances and Savings Plans deliver lower cost than on-demand pricing.

Per-choice explanations

Practical Use

Reserved Instances (RI) and Savings Plans both provide discounts off on-demand pricing in exchange for a one- or three-year commitment. RIs are purchased per service (EC2, RDS, ElastiCache, Redshift, and others) and specify the instance type, Region, OS, and similar attributes. Savings Plans commit to spending a fixed dollar amount per hour, and Compute Savings Plans apply flexibly across EC2, Fargate, and Lambda. For workloads that run steadily, savings can reach roughly 72% compared with on-demand. As a rule of thumb, choose Compute Savings Plans (higher flexibility) when the instance types or services you use may change, and choose RIs when the configuration is fixed, such as a specific DB instance or Redshift cluster. RIs come in Standard RI (higher discount, limited changes) and Convertible RI (slightly lower discount, instance type changes allowed). Selecting a three-year term with all upfront payment yields the maximum discount. Spot Instances, by contrast, suit batch and asynchronous workloads that can tolerate interruption, and are not appropriate on their own for steady workloads. On SAA-C03, conditions such as "continuous operation," "usage confirmed for at least one year," and "reduce cost compared to on-demand" point to RI or Savings Plans. Distractors such as Spot only, instance store, and additional Elastic IPs either conflict with the stable workload requirement or are not cost-reduction techniques. Commitment terms (one vs. three years), payment options (all upfront, partial upfront, no upfront), and the difference between Standard and Convertible RI are also asked in higher-level exams.