Examkin

Soru

Zorluk: Çok zorCost-Optimized Database Design and Capacity Planning

A logistics organization tracks real-time location telemetry for its fleet of vehicles. The fleet operations generate a highly predictable write volume: from Monday through Friday between 08:00 and 18:00, telemetry ingestion averages 15,00015,000 write requests per second, whereas overnight and during weekends, it drops to a baseline of 100100 write requests per second. Additionally, once a month, an internal audit team executes a heavy read-intensive batch job that scans the entire database to generate compliance reports, which takes approximately 2 hours. Which TWO of the following configurations represent the most cost-effective database capacity design to meet these requirements?

  1. Configure the Amazon DynamoDB table in Provisioned Capacity Mode and implement Application Auto Scaling with scheduled actions to scale up Write Capacity Units (WCUs) to 15,00015,000 before 08:00 on weekdays and scale down to 100100 WCUs during off-peak hours.Cevap
  2. Enable Point-in-Time Recovery (PITR) on the DynamoDB table, export the table data directly to Amazon S3 once a month, and perform the compliance audit queries using Amazon Athena.Cevap
  3. C
    Configure the Amazon DynamoDB table in On-Demand Capacity Mode to dynamically accommodate the 15,00015,000 writes per second during weekdays and automatically scale down to 100100 writes per second overnight.
  4. D
    Configure the database on an Amazon RDS Multi-AZ DB instance deployment, and configure the monthly compliance audit job to read directly from the passive standby DB instance to avoid performance impact on the primary instance.
  5. E
    Export the DynamoDB table data to Amazon S3 Standard-Infrequent Access (S3 Standard-IA), execute the compliance queries using Amazon Athena, and immediately delete the exported data in S3 after the 2-hour audit completes to minimize storage costs.

Cevap

Configure the DynamoDB table in Provisioned Capacity Mode with scheduled scaling to transition between 15,00015,000 and 100100 WCUs, and export the table data via Point-in-Time Recovery (PITR) to Amazon S3 to query using Amazon Athena.
The correct database design incorporates Provisioned Capacity Mode with scheduled scaling for writes, combined with PITR exports to Amazon S3 and Amazon Athena for read-heavy compliance audits. Scheduled scaling aligns capacity with predictable changes, reducing write costs by over 85% compared to On-Demand capacity mode. Exporting DynamoDB table data using PITR does not consume any live table capacity, eliminating the need to provision high Read Capacity Units (RCUs) for monthly 2-hour compliance scans.

Adım Adım Çözüm

1
Analyze the write workload characteristics and predictability.
The write workload has a predictable, high-volume peak of 15,00015,000 writes/sec during weekdays and drops to 100100 writes/sec overnight and on weekends.
Understanding whether traffic is predictable is key to choosing between DynamoDB On-Demand and Provisioned Capacity modes.
2
Calculate the comparative costs of On-Demand vs. Provisioned Capacity with Scheduled Scaling.
On-Demand write costs: 15,000 writes/sec×3600 sec/hour×10 hours/day×22 days/month=11.88 billion writes/month15,000 \text{ writes/sec} \times 3600 \text{ sec/hour} \times 10 \text{ hours/day} \times 22 \text{ days/month} = 11.88 \text{ billion writes/month}. At $1.25\$1.25 per million, this costs $14,850.00\$14,850.00 per month. Provisioned capacity costs: 15,000 WCUs×220 hours×$0.00065≈$2,145.0015,000 \text{ WCUs} \times 220 \text{ hours} \times \$0.00065 \approx \$2,145.00 + off-peak baseline costs of $33.15\$33.15, totaling $2,178.15\$2,178.15 per month.
Determining the most cost-effective capacity option for the database writes.
3
Evaluate options for offloading the monthly 2-hour database scan.
Scanning the live database directly requires provisioning massive Read Capacity Units (RCUs) which is highly expensive and risks throttling. Exporting the database using PITR to S3 does not consume live read capacity and allows querying via Amazon Athena for a fraction of the cost.
Decoupling analytical queries from transactional workloads to optimize cost and performance.

Anahtar Kavram

Using scheduled scaling for predictable capacity adjustments and offloading read-heavy analytical scans using serverless query engines on exported data to minimize total database costs.
Tahmini Süre:3m 0s
Bu soruyu puanla

Konular

Cost-Optimized Database Design and Capacity Planning