Soru

Zorluk: Çok zorDatabase and Storage Strategy

A financial institution is designing a new global OLTP platform to process trades across a primary region (useast1us-east-1) and a secondary recovery region (euwest1eu-west-1). The architecture must support database writes with a Recovery Point Objective (RPO) of less than 5 seconds and a Recovery Time Objective (RTO) of less than 1 minute. Microsecond-latency read caching is required locally in both regions. Security compliance dictates that all data at rest must be encrypted using customer-managed keys (CMKs) to enable cross-account access for centralized auditing. Additionally, the system must dynamically scale read capacity in the secondary region during sudden traffic spikes without manual intervention. Which two actions should the solutions architect take to meet these database and storage requirements?

  1. Deploy an Amazon Aurora PostgreSQL Global Database with the primary cluster in the primary region and a secondary cluster in the secondary region, encrypting both clusters with region-specific KMS Customer Managed Keys, and configure Aurora Auto Scaling on Replicas in the secondary region.Cevap
  2. Deploy an Amazon ElastiCache for Redis Global Datastore with the primary cluster in the primary region and a secondary cluster in the secondary region to provide local microsecond-latency caching.Cevap
  3. C
    Deploy an Amazon Aurora PostgreSQL Global Database using the default AWS-managed KMS key (aws/rds) to encrypt the clusters, and create a cross-account IAM policy to grant key access to the centralized auditing account.
  4. D
    Configure Amazon RDS PostgreSQL Multi-AZ standby instances in the secondary region to serve as the read-scaling endpoints during traffic spikes by modifying the DB parameter group.
  5. E
    Deploy Amazon ElastiCache for Memcached in a multi-region cluster configuration to cache query results with global data replication.

Cevap

Deploy an Amazon Aurora PostgreSQL Global Database using region-specific Customer Managed Keys (CMKs) with Auto Scaling configured on the secondary replicas, and deploy an Amazon ElastiCache for Redis Global Datastore for multi-region microsecond read caching.
The correct architecture requires a combination of Amazon Aurora PostgreSQL Global Database (using Customer Managed Keys for cross-account compliance and Aurora Replicas for dynamic read scaling) and Amazon ElastiCache for Redis Global Datastore (for local microsecond-latency caching across regions). This setup meets all performance, HA/DR, security, and scalability requirements.

Adım Adım Çözüm

1
Select the database engine that meets the global transaction and recovery constraints.
Amazon Aurora Global Database is chosen because it replicates data with a typical latency of less than 1 second, fulfilling the RPO (< 5 seconds) and RTO (< 1 minute) parameters.
Standard multi-region RDS read replicas do not offer the same low-latency replication or rapid failover capabilities as Aurora Global Databases.
2
Determine the correct KMS encryption key type for cross-account compliance.
Customer Managed Keys (CMKs) must be configured in each region for the Aurora cluster.
AWS-managed KMS keys cannot have their key policies edited, preventing cross-account access delegation to the security/audit account.
3
Establish the global caching layer.
Amazon ElastiCache for Redis Global Datastore is deployed across both regions.
ElastiCache for Redis supports cross-region replication and provides local, microsecond-latency reads, whereas Memcached lacks replication and global datastore functionality.
4
Configure scaling policies for the database read tier.
Aurora Auto Scaling is enabled on the Replicas in the secondary region.
This automatically scales the number of Aurora Replicas up or down based on CPU utilization or average connections, whereas Multi-AZ standby instances cannot serve read traffic or scale dynamically.

Anahtar Kavram

Designing multi-region database and storage architectures with strict RTO/RPO limits, microsecond caching requirements, custom encryption key policies, and dynamic read scaling.
Bu soruyu puanla