For which use case are Amazon EC2 On-Demand Instances MOST cost-effective?

Question

Options

• ACompute-intensive video transcoding that can be restarted if necessary
• BAn instance in continual use for 1 month to conduct quality assurance tests
• CAn instance that runs a web server that will run for 1 year
• DAn instance that runs a database that will run for 3 years

• A
  4% (2)
• B
  76% (34)
• C
  7% (3)
• D
  13% (6)

Explanation

On-Demand Instances charge per hour/second with no upfront commitment, making them ideal for short-term, unpredictable workloads - which is exactly what option B describes: a one-month QA test with continuous usage that doesn't justify a longer commitment.

Why the distractors are wrong:

• A describes an interruptible workload, which is the textbook use case for Spot Instances (up to 90% cheaper, with the tradeoff of potential interruption).
• C (1-year web server) and D (3-year database) both involve predictable, long-running workloads - perfect for Reserved Instances, which offer significant discounts (up to 72%) in exchange for a 1- or 3-year commitment.

Memory tip: Match the pricing model to the workload duration - think of On-Demand as a "hotel room" (flexible, no contract, pays per night), Spot as a "last-minute deal" (cheap but can be cancelled), and Reserved as signing a "lease" (commit long-term, pay less per month).

No community discussion yet for this question.