Question

Difficulty: MediumUnderstand concepts of cloud economics

VividArchive is migrating its video digitization and rendering platform to AWS. The platform has two primary workloads:

1. A continuous, predictable digitization process that runs 24/7 to archive historical tapes.
2. Sudden, unpredictable requests from news agencies for high-definition rendering, which requires massive compute capacity for brief periods.

Which two strategies will help the company minimize its Total Cost of Ownership (TCO) by aligning AWS billing models with these workload characteristics? (Select TWO.)

  1. Deploying the digitization workload on Amazon EC2 instances covered by a Savings Plan to benefit from commitment-based discountsAnswer
  2. Scaling the rendering capacity dynamically using Amazon EC2 Auto Scaling and On-Demand instances to match peak demandsAnswer
  3. C
    Purchasing On-Demand instances with a five-year upfront payment to secure the lowest rates for the unpredictable rendering spikes
  4. D
    Tightly coupling the digitization and rendering services onto a single high-capacity EC2 instance to minimize cross-service networking fees
  5. E
    Provisioning a fixed fleet of instances sized to the historical peak demand of rendering requests to guarantee zero latency

Answer

Deploying the digitization workload on Amazon EC2 instances covered by a Savings Plan, and scaling the rendering capacity dynamically using Amazon EC2 Auto Scaling and On-Demand instances.
The correct strategies leverage the core cloud economics principles of elasticity and commitment-based pricing. The baseline, predictable tape digitization workload runs continuously, which makes it a perfect candidate for Savings Plans or Reserved Instances to lower the cost per hour. Conversely, the unpredictable spikes in video rendering demands are best handled through dynamic scaling using On-Demand instances, ensuring the company only pays for compute resources when they are actively processing requests.

Step-by-Step Solution

1
Analyze the resource consumption pattern of the tape digitization process.
Identified as a continuous, baseline workload that runs 24/7 with high predictability.
Baseline workloads are best optimized using commitment-based pricing models like Savings Plans or Reserved Instances to lower the hourly rate.
2
Analyze the resource consumption pattern of the video rendering process.
Identified as a highly spikey, unpredictable, and short-duration workload.
Spikey workloads should utilize elasticity through Auto Scaling and On-Demand instances to scale down when not in use, avoiding paying for idle resources.

Key Concept

Matching pricing models and elasticity to workload predictability
Estimated Time:1m 30s
Rate this question