Examkin

Soru

Zorluk: Çok zorDesigning Infrastructure for Technical Requirements and High Availability

An enterprise financial trading platform is being re-architected on Google Cloud to satisfy critical business continuity constraints: a Recovery Point Objective (RPO) of 00 (zero data loss) and a Recovery Time Objective (RTO) of less than 11 minute in the event of a full regional infrastructure outage. The architecture must also support high-throughput, low-latency relational data access and isolated network communication across multiple internal Virtual Private Cloud (VPC) networks without introducing transitive routing anti-patterns or cascading dependency failures under heavy load. Which TWO architectural decisions should you implement to meet these high availability and technical requirements?

  1. Provision a multi-region Cloud Spanner instance configuration spanning multiple geographic regions to serve as the core transactional database layer.Cevap
  2. Deploy regional application workloads in separate GCP regions behind a Global External Application Load Balancer using Serverless NEGs and automated multi-region health probes.Cevap
  3. C
    Provision a Highly Available (HA) regional Cloud SQL instance with cross-region asynchronous read replicas, promoting a read replica during regional disaster recovery.
  4. D
    Connect spoke VPC networks to a central hub VPC using VPC Network Peering, enabling transitive network communication across all spoke VPCs via the central hub.
  5. E
    Configure Load Balancer health checks to execute complex database query endpoints directly to guarantee backend application data layer availability.

Cevap

Deploying a multi-region Cloud Spanner instance along with multi-region compute behind a Global External Application Load Balancer satisfies the zero RPO and sub-minute RTO availability requirements.
Combining a multi-region Cloud Spanner instance with multi-region compute services behind a Global External Application Load Balancer guarantees multi-region resilience. Cloud Spanner uses Paxos-based synchronous cross-region replication to ensure zero data loss (RPO=0RPO = 0), while global load balancing provides sub-minute traffic failover (RTO<1 minRTO < 1\text{ min}) upon regional failure.

Adım Adım Çözüm

1
Evaluate database requirements for RPO=0 and high availability across regional outages.
Cloud Spanner multi-region configurations utilize Paxos consensus across regional zones to deliver synchronous cross-region write replication, meeting RPO=0 and high availability SLAs.
Cloud SQL cross-region read replicas use asynchronous replication, which violates the strict RPO=0 requirement.
2
Evaluate compute routing and global load balancing strategy.
Deploying global HTTP(S) load balancing across multi-region backend compute endpoints allows instant health-based traffic redirection without DNS propagation delays.
Global load balancing delivers RTO under 1 minute for application tier failover.
3
Verify network isolation and health check constraints.
Avoid non-transitive VPC peering assumptions and deep dependency health check probes.
VPC peering does not support transitive routing, and deep health checks introduce risk of cascading cluster outages.

Anahtar Kavram

High Availability and Multi-Region Resiliency Architecture
Bu soruyu puanla

Konular

Designing Infrastructure for Technical Requirements and High Availability