SAA-C03 Question 499
Single answerYour company is deploying a new e-commerce platform, and the team has decided to use Amazon DynamoDB to store customer orders. The workload is expected to have highly variable traffic patterns, with spikes during promotional events. You want to ensure the database can scale to handle peak loads while minimizing costs during periods of low activity. What is the most appropriate capacity planning configuration for this use case?
- A
Provisioned capacity with sufficient read and write capacity units to handle peak traffic.
- B
DynamoDB On-Demand capacity mode to automatically scale with traffic without pre-provisioning.
- C
Enable Auto Scaling on provisioned capacity to dynamically adjust capacity based on traffic.
- D
Use Reserved Capacity for predictable traffic patterns to save costs.
Show answer and explanation
Correct answer: B
Explanation
DynamoDB On-Demand capacity mode is the best choice for workloads with unpredictable and highly variable traffic patterns, as it eliminates the need for manual capacity planning or pre-provisioning of capacity units. It automatically scales up or down based on traffic, optimizing both performance and cost-efficiency.
- A. Incorrect.
Provisioned capacity requires you to manually estimate and allocate capacity units for peak traffic, which can lead to over-provisioning during low activity periods or under-provisioning during traffic spikes. This is not cost-effective or suitable for highly variable workloads.
- B. Correct.
DynamoDB On-Demand capacity mode is ideal for workloads with unpredictable or highly variable traffic patterns. It automatically adjusts to traffic, scaling up and down as needed, ensuring cost-efficiency and availability during both peak and low usage periods.
- C. Incorrect.
Auto Scaling on provisioned capacity can adjust capacity dynamically, but it requires configuring scaling policies and pre-allocating a base capacity, which might still result in inefficiencies for highly variable traffic patterns.
- D. Incorrect.
Reserved Capacity is a cost-saving option for predictable traffic patterns, but it is not suitable for workloads with unpredictable or spiky traffic.