Design for New Solutions IV
Explore designing robust AWS architectures for global applications, high-performance computing, and serverless analytics. Learn to choose appropriate databases, optimize latency with multi-Region strategies, manage concurrent file access, and implement cost-efficient compute models. Understand trade-offs to build resilient, scalable cloud systems that remain available under failure conditions.
We'll cover the following...
Question 36
A travel company is building a global shopping cart service. Users in North America and Europe must read and write locally with single-digit millisecond latency, and the service must remain available if one Region fails. The data model is key-value and can tolerate last-writer-wins conflict handling with application safeguards. Which design is best?
A. Use Amazon Aurora Global Database with one writable primary Region and a secondary Region, and route all writes to the primary Region to preserve relational consistency.
B. Deploy stateless compute in each Region, use DynamoDB global tables for active-active Regional reads and writes with asynchronous replication, design high-cardinality partition keys, use conditional writes or idempotency keys to handle duplicates and conflicts, and route users with Amazon Route 53 latency-based routing or AWS Global Accelerator.
C. Store cart state as objects in Amazon S3 with Cross-Region Replication, and use Amazon Route 53 simple routing to send all users to the closest object replica.
D. Use Amazon RDS Multi-AZ in one Region with read replicas in the second Region, and rely on Multi-AZ failover for Regional survivability.
Question 37
A social application needs a global user preference and session metadata store. Access is key-value, traffic is ...