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.

• ✗

  Pilot light

  Why it's wrong here

  The 'pilot light' strategy replicates only the most critical data (e.g., database snapshots or an RDS cross-region read replica) while keeping a minimal set of infrastructure ready in the secondary Region. On failover, you must manually or programmatically launch EC2 instances from pre-baked AMIs and configure the environment, which normally consumes minutes to hours. Since the data replication is typically asynchronous (e.g., S3 CRR, snapshot shipping or an Aurora global database with asynchronous replication), the RPO is not near-zero; it is equal to the lag between the last replicated transaction and the failure. The pilot light is cost-efficient but simply cannot meet strict recovery targets requiring immediate switchover with no data loss.

• ✗

  Backup and restore

  Why it's wrong here

  Backup and restore is the simplest DR pattern, where you periodically snapshot EBS volumes, RDS databases, AMIs, or S3 data and, in a disaster, restore them into a new Region or VPC. The RPO is bounded by the backup scheduling interval—anything written after the last backup is lost—and for high-write workloads that gap can represent hours or even days of data. The RTO is also the longest of all patterns because you must rehydrate entire AMIs, attach EBS volumes, possibly re-subscribe to events, and wait for data transfers over the network; this can take hours, if not days, for large datasets. While backups are the cheapest option (no idle compute), they offer the weakest recovery guarantees and are only suitable when tolerating significant data loss and downtime is acceptable.

• ✗

  Warm standby

  Why it's wrong here

  Warm standby keeps a fully functional but scaled-down copy of the production environment running in a secondary Region, including a mirrored database, an Auto Scaling group, and the same application stack. Because the compute and networking resources are already live, failover can be completed in minutes by scaling up capacity and redirecting DNS (e.g., through Route 53 health checks) without provisioning time. The tradeoff is that the standby infrastructure incurs continuous compute and storage spend, making it costlier than pilot light or backups. Additionally, if the data replication is asynchronous, the RPO may be a few seconds to minutes of lag, so it does not guarantee zero loss—and because the standby is not actively serving traffic, you must still manage the state and connection draining during cutover.

• ✓

  Multi-Region active-active

  Why this is correct

  A true multi-Region active-active design has production workloads live in two or more AWS Regions simultaneously, with traffic typically distributed by Route 53 latency routing, weighted routing, or an AWS Global Accelerator. For this pattern to achieve low RPO, data synchronization must be synchronous or effectively near-synchronous, as seen with Amazon DynamoDB global tables or Aurora Global Database with write forwarding. Because application instances are already running and ready to serve requests in every Region, failover can be nearly instantaneous—RTO can be sub-minute—and with multi-master synchronous writes, the RPO can be zero. This architecture demands the highest complexity and cost, requiring cross-region conflict resolution, careful management of session affinity, and application-level multi-region awareness; it is the only option that meets strict recovery point and recovery time objectives.

Go deeper

Related to this question

Same concept, more angles

1 more way this is tested on DOP-C02

These questions test the same concept from different angles. Work through them to make sure you can recognise it however the exam phrases it.