Post

Q22: Storing Media Files With Unpredictable Access Patterns at Minimum Cost

A digital media app's files must survive the loss of an Availability Zone, with some files accessed frequently and others rarely in an unpredictable pattern, minimizing storage and retrieval costs.

Q22: Storing Media Files With Unpredictable Access Patterns at Minimum Cost

Problem Info

  
DomainDesign Cost-Optimized Architectures
ServicesAmazon S3, S3 Intelligent-Tiering
DifficultyEasy–Medium

Question

A digital media app stores files in S3. Files must be resilient to the loss of an Availability Zone. Some files are accessed frequently; others are rarely accessed in an unpredictable pattern. The architect must minimize storage and retrieval costs.

Which storage option meets these requirements?

Correct Answer

S3 Intelligent-Tiering.

Why this is correct

The phrase “unpredictable access pattern” is the giveaway here — it’s the exact scenario S3 Intelligent-Tiering was built to solve.

Every S3 storage class (Standard, Standard-IA, Intelligent-Tiering, One Zone-IA, Glacier tiers) that spans multiple Availability Zones by default satisfies “resilient to the loss of an AZ” — that rules out One Zone-IA immediately, but leaves several candidates. What separates Intelligent-Tiering is its automatic cost optimization: it monitors each object’s access pattern and automatically moves objects between a frequent access tier (priced like S3 Standard) and an infrequent access tier (priced lower), with no retrieval fees for moving between these tiers, and no operational effort required to classify files yourself.

Because the files in this scenario are accessed unpredictably — sometimes frequently, sometimes not, with no clear schedule — manually choosing between Standard and Standard-IA (or writing custom Lifecycle rules based on a fixed number of days) risks either overpaying for Standard when a file goes cold, or incurring early-retrieval penalty fees on Standard-IA if a “cold” file is accessed sooner than expected. Intelligent-Tiering removes that guesswork and those retrieval-fee risks entirely.

Why the alternatives fall short

  • S3 Standard for everything — resilient to AZ loss, but you pay the highest per-GB rate even for files that go untouched for months — fails to minimize cost.
  • S3 Standard-IA with a Lifecycle policy based on a fixed access window — cheaper for genuinely infrequent files, but Standard-IA charges a retrieval fee every time a file is read; with an unpredictable pattern, some “infrequent” files may be reread often, generating unexpected retrieval costs. Lifecycle rules also assume a known time-based pattern, which doesn’t fit “unpredictable.”
  • S3 One Zone-IA — cheaper than Standard-IA, but stores data in only a single Availability Zone, directly violating “resilient to the loss of an Availability Zone.”

Exam Tip

“Unpredictable / unknown / changing access patterns” + “minimize cost” → S3 Intelligent-Tiering, almost without exception on this exam. Reserve Lifecycle policies to Standard-IA/Glacier tiers for scenarios where the access pattern is known and predictable (e.g., “accessed frequently for 30 days, then rarely” — see Q23), and always remember One Zone-IA sacrifices AZ resilience for a lower price.

This post is licensed under CC BY 4.0 by the author.