Examkin

Question

Difficulty: EasyHigh-Performing Database Solutions

A solutions architect is designing a database tier for a new high-throughput social networking application. The application requires microsecond read latency to serve a small list of popular trending profiles. The write volume is expected to scale up to 15,00015,000 transactions per second, and the architect must ensure that these write operations are evenly distributed to prevent partition bottlenecks on the underlying Amazon DynamoDB table.

Which TWO options should the solutions architect select to meet these performance requirements?

  1. Deploy an Amazon DynamoDB Accelerator (DAX) cluster to cache read requests for popular profiles.Answer
  2. Choose a high-cardinality attribute, such as a unique UserID, as the partition key.Answer
  3. C
    Choose a monotonically increasing attribute, such as a registration timestamp, as the partition key.
  4. D
    Configure Amazon RDS Read Replicas to act as the primary write target during peak workloads.

Answer

Deploying an Amazon DynamoDB Accelerator (DAX) cluster to cache read requests for popular profiles, and choosing a high-cardinality attribute, such as a unique UserID, as the partition key.
The correct architecture requires deploying a DynamoDB Accelerator (DAX) cluster to provide in-memory microsecond read responses for the trending profiles. It also requires selecting a high-cardinality partition key, such as UserID, to distribute the write load of 15,00015,000 transactions per second evenly across partitions and avoid write bottlenecks.

Step-by-Step Solution

1
Evaluate read latency requirements
The application requires microsecond latency for hot read requests, indicating the need for an in-memory cache rather than direct database queries.
DynamoDB Accelerator (DAX) provides microsecond latency for read operations on DynamoDB tables.
2
Evaluate write distribution requirements
To support 15,00015,000 transactions per second without partition limits being exceeded, the database must distribute writes uniformly.
Choosing a high-cardinality partition key like UserID prevents hot partitions by spreading writes across multiple physical partitions.

Key Concept

High-performing DynamoDB architectures rely on in-memory caching (DAX) for sub-millisecond read access and high-cardinality partition keys to distribute read/write throughput uniformly across partitions.
Estimated Time:1m 30s
Rate this question

Topics

High-Performing Database Solutions

More questions from this topic