DEA-C01 Practice Question 24: S3 Intelligent Tiering Unknown Access
AWS Certified Data Engineer – Associate (DEA-C01) · Test 3 · question 4 of 10
A curated data lake zone holds 400 TB of Parquet. Access is genuinely unpredictable: a dataset may be queried heavily for three weeks while a model is being built, then ignored for months, and analysts occasionally query two-year-old data and expect it back immediately. Storage cost has become the largest line item.
The team wants to cut storage cost without retrieval fees, without any restore step before a query, and without building a usage-tracking process.
Which storage configuration should they use?
Show answer and explanation
Answer:
S3 Intelligent-Tiering is the storage class documented for "data with unknown, changing, or unpredictable access patterns". For a small monthly monitoring and automation fee per object, it moves objects between the Frequent Access, Infrequent Access (30 days without access) and Archive Instant Access (90 days without access) tiers automatically, and moves them straight back to Frequent Access on the next access. Critically for this scenario, all three of those tiers are low-latency, and Intelligent-Tiering has no retrieval fees, so the occasional query against two-year-old data costs nothing extra and needs no restore. The optional Archive Access and Deep Archive Access tiers are opt-in, so leaving them off keeps everything instantly readable.
Option A charges a per-GB retrieval fee on every read. The dataset that gets queried heavily during a three-week modelling push would be read repeatedly out of Standard-IA, and the retrieval fees can easily exceed the storage savings. Standard-IA also has a 30-day minimum storage duration.
Option B breaks the immediacy requirement outright. S3 Glacier Flexible Retrieval objects "are archived, and not available for real-time access"; you must restore an object before reading it, which takes minutes to hours. It also carries a 90-day minimum duration and retrieval fees.
Option C is cheaper than Standard-IA precisely because it stores one copy in a single Availability Zone, so "the data is not resilient to the physical loss of the Availability Zone." For a curated lake zone that is the primary copy, that is the wrong trade, and it still has retrieval fees.
Reference: Understanding and managing Amazon S3 storage classes