A global travel booking company runs a relational booking system on AWS. The application h… — AWS Certified Solutions Architect - Associate

A global travel booking company runs a relational booking system on AWS. The application has its primary deployment in the `us-east-1` region but serves a growing user base in `eu-west-1` and `ap-southeast-1`. The database workload consists of a $90:10$ read-to-write ratio, and the system must handle peak write volumes of up to $10,000$ transactions per second. To improve performance and resilience, a solutions architect must design a database architecture that achieves:

* Low-latency local read access (under $30\text{ ms}$ ) in all three regions.
* Minimal replication lag and zero performance overhead on the primary writer during peak write volumes.
* A disaster recovery failover capability to another region with a Recovery Time Objective (RTO) of less than $1\text{ minute}$ .

Which database configuration should the solutions architect recommend to meet these requirements?

Deploy Amazon Aurora Global Database with the primary cluster in `us-east-1` and secondary clusters in `eu-west-1` and `ap-southeast-1`. Configure local application instances to read from their local Aurora replica instances, and use Aurora storage-level replication to keep data synchronized.Answer
B
Deploy Amazon RDS for PostgreSQL in `us-east-1` with Multi-AZ enabled, and create cross-region read replicas in `eu-west-1` and `ap-southeast-1`. Configure the application to route read traffic to the local replicas and perform automatic failover to the replicas using Amazon Route 53.
C
Migrate the booking schema to Amazon DynamoDB with Global Tables across the three regions. Use the booking date (`YYYY-MM-DD`) as the partition key to store and retrieve customer reservation details.
D
Migrate the database to Amazon DynamoDB Global Tables across the three regions, configuring the tables in provisioned capacity mode with fixed read and write units to handle average traffic levels and avoid the latency overhead of auto-scaling.

Answer

Deploy Amazon Aurora Global Database with the primary cluster in `us-east-1` and secondary clusters in `eu-west-1` and `ap-southeast-1`. Configure local application instances to read from their local Aurora replica instances, and use Aurora storage-level replication to keep data synchronized.

The correct configuration uses Amazon Aurora Global Database because its replication is handled at the storage layer using dedicated infrastructure rather than the database engine, ensuring that replication tasks do not impact the write performance of the primary instance in `us-east-1` even during peak volumes of

10,000

transactions per second. This storage-level replication maintains cross-region replication lag under

1\text{ second}

and supports RTO of less than

1\text{ minute}

when promoting a secondary region cluster to a primary read-write cluster. It also natively supports the relational booking schema.

Step-by-Step Solution

Analyze the database schema and application requirements.

The application requires a relational database schema. This eliminates NoSQL solutions like Amazon DynamoDB from the correct choices, although we must also evaluate the performance characteristics of the remaining relational databases.

Ensuring the selected database tier supports relational SQL transactional integrity with ACID compliance.

Evaluate replication performance under heavy write volumes.

Amazon RDS uses logical engine-level replication which creates performance overhead on the writer node and increases replication lag during peaks. Amazon Aurora Global Database uses storage-level replication, separating replication tasks from the primary writer node to guarantee minimal replication lag (under

1\text{ second}

) without performance impact.

Meeting the performance requirement of zero write overhead on the primary instance under a load of

10,000

transactions per second.

Evaluate cross-region disaster recovery promotion times (RTO).

In the event of a regional outage, an Aurora Global Database secondary region can be promoted to primary in less than

1\text{ minute}

. Promoting an RDS cross-region read replica is manual and takes several minutes.

Checking compliance with the Disaster Recovery RTO requirement of less than

1\text{ minute}

Key Concept

Leveraging Amazon Aurora Global Database storage-level replication for global low-latency reads, minimal write performance overhead, and fast RTO cross-region failover.

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic