DOP-C02 Resilient Cloud Solutions 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.

• ✗

  DynamoDB point-in-time recovery (PITR)

  Why it's wrong here

  DynamoDB point-in-time recovery (PITR) continuously backs up a table within a single Region, allowing restores to any second within the last 35 days. However, it is a backup/restore mechanism, not an availability or failover feature: if the Region becomes unavailable, PITR cannot redirect traffic or serve reads/writes. You must manually restore to a table (often in a different Region, if you have set that up), and that restore operation takes time, so it does not meet low-RTO disaster recovery requirements.

• ✓

  DynamoDB global tables

  Why this is correct

  DynamoDB global tables create a fully managed, multi-Region, multi-master replicated database using DynamoDB Streams and a backend replication engine. Changes made in any replica table are replicated to all other selected Regions with sub-second latency, enabling active-active failover: if one Region fails, traffic can be redirected to a remaining Region with minimal or zero downtime. This architecture directly addresses the requirement for low-latency access and resilience to Regional outages, making it the correct answer.

• ✗

  DynamoDB Accelerator (DAX)

  Why it's wrong here

  DynamoDB Accelerator (DAX) is an in-memory cache that provides microsecond response times for individual table requests, but it operates entirely within a single Region and is tightly coupled to the same VPC and endpoint. It does not replicate data across Regions, nor does it provide any standby or failover capability for a Regional failure. While it can improve read performance, it has no role in disaster recovery and would be unavailable if the underlying Region failed.

• ✗

  DynamoDB on-demand capacity mode

  Why it's wrong here

  DynamoDB on-demand capacity mode automatically scales read/write throughput to match actual traffic patterns, charging per request instead of for provisioned capacity. It is a scaling and billing model that helps absorb sudden spikes without capacity planning, but it does not move data, replicate tables, or provide any cross-Region redundancy. Enabling on-demand mode has no effect on data durability or availability during a Regional outage, so it is incorrect for this disaster recovery scenario.

Go deeper

Related to this question