Question #841
A company is launching a new video streaming service on Amazon EC2 instances. The service must be available globally across three AWS Regions: us-east-1, eu-west-1, and ap-southeast-1. User preferences, watch history, and session data must be accessible across all Regions. A solutions architect needs to design a solution allowing any Region to scale to handle the load of all Regions. Additionally, users must automatically connect to the Region with the least latency. Which solution meets these requirements with the LEAST operational overhead?
Use EC2 Spot Fleets with Network Load Balancers in each Region. Configure AWS Global Accelerator to point to the NLBs. Use Route 53 latency-based routing for the Global Accelerator IP. Store data in Amazon RDS for PostgreSQL with read replicas in other Regions.
Deploy Auto Scaling groups and NLBs in each Region. Use Route 53 geoproximity routing pointing to each NLB. Store data in self-managed PostgreSQL on EC2 instances with cross-region replication.
Deploy Auto Scaling groups with NLBs in each Region. Use Route 53 latency-based routing pointing to each NLB. Store data in Amazon DynamoDB global tables.
Use EC2 instances in each Region with NLBs. Deploy a custom DNS solution on EC2 instances using latency checks to route users. Store data in Amazon Aurora Global Database.
Explanation
Option C meets all requirements effectively:
1. Latency-Based Routing: Route 53 directs users to the nearest Region, ensuring minimal latency.
2. Scalability: Auto Scaling groups allow each Region to handle increased load dynamically.
3. Cross-Region Data Access: DynamoDB global tables provide multi-region replication with read/write capabilities, ensuring data consistency globally.
4. Operational Efficiency: Managed services (Route 53, DynamoDB, Auto Scaling) minimize operational overhead.
Other options fail due to:
- A: RDS read replicas are read-only, limiting cross-region writes. Global Accelerator + Route 53 is redundant.
- B: Self-managed PostgreSQL increases operational effort; geoproximity routing is less precise.
- D: Aurora Global Database lacks multi-master writes; custom DNS adds complexity.
Key Points:
- Use managed services (DynamoDB global tables) for cross-region data.
- Route 53 latency-based routing optimizes user connections.
- Auto Scaling ensures regional scalability.
Answer
The correct answer is: C