Examkin

Question

Difficulty: HardPerformance and Scalability Optimization

A digital advertising company is designing a real-time bidding (RTB) platform on AWS. The platform must ingest and process 2.5 million2.5\text{ million} bid requests per second from external ad exchanges. Each request has a strict latency budget of 50 ms50\text{ ms} for processing, matching with active advertiser campaigns, and returning a bid response. The campaign metadata is currently 15 GB15\text{ GB} in size and is updated periodically throughout the day. The architecture must scale horizontally to handle unpredictable spikes in bid requests while maintaining the sub-50 ms50\text{ ms} response time. Which combination of actions should the Solutions Architect take to meet these performance and scalability requirements? (Select TWO.)

  1. Cache the campaign metadata using Amazon ElastiCache for Redis with cluster mode enabled, and configure the application to read from replica nodes to minimize latency and distribute the query volume.Answer
  2. B
    Deploy an Application Load Balancer (ALB) to manage incoming requests, relying on its dynamic auto-scaling to scale the load balancer's capacity without pre-warming during sudden surges.
  3. Deploy a Network Load Balancer (NLB) to handle incoming traffic, terminating TCP/TLS connections at the load balancer to reduce computing load on target instances.Answer
  4. D
    Store the campaign metadata in an Amazon RDS MySQL database in a Multi-AZ deployment, and configure the application instances to direct read traffic to the standby replica in the second Availability Zone to scale out reads.
  5. E
    Deploy Amazon ElastiCache for Memcached to store the campaign metadata, and configure multi-AZ replication to ensure the cache stays highly available and in-sync across multiple Availability Zones.

Answer

Deploy a Network Load Balancer (NLB) to handle incoming traffic, terminating TCP/TLS connections at the load balancer, and cache the campaign metadata using Amazon ElastiCache for Redis with cluster mode enabled, reading from replica nodes to minimize latency.
Deploying a Network Load Balancer (NLB) to handle traffic is correct because it scales instantly to support millions of requests per second with ultra-low latency and offloads TLS termination to preserve computing capacity on the application tier. Storing campaign metadata in Amazon ElastiCache for Redis with cluster mode enabled is correct because it fits the 15 GB15\text{ GB} dataset in-memory and supports multi-AZ read replication, allowing application instances to execute queries against local replica nodes with sub-millisecond response times.

Step-by-Step Solution

1
Evaluate the ingestion mechanism for extreme throughput and latency requirements.
Using a Network Load Balancer (NLB) is selected because it operates at the transport layer (Layer 4) and is built to handle millions of requests per second with sub-millisecond latencies, scaling instantly without the pre-warming delays inherent in Application Load Balancers.
An Application Load Balancer would drop requests during instant, massive spikes because it relies on a scaling process that cannot keep pace with flash traffic.
2
Determine the optimal caching layer for rapid, highly available metadata retrieval.
Amazon ElastiCache for Redis with cluster mode enabled is selected because the 15 GB15\text{ GB} dataset easily fits in memory, and Redis supports read replication across multiple Availability Zones to scale read operations sub-millisecond.
Memcached does not support multi-AZ replication to satisfy high availability constraints, and RDS Multi-AZ standby instances are passive and cannot serve reads.

Key Concept

Decoupling high-throughput load balancing and ultra-low latency caching to scale workloads under strict sub-second performance SLAs.
Rate this question

Topics

Performance and Scalability Optimization

More questions from this topic