PDE - Storing Data (20% of the exam) - Section 3.3

Manage data lakes by configuring discovery, access controls, cost controls, data processing, and monitoring.

Configure a Cloud Storage-backed data lake using Dataplex for automated data discovery, access control policies, and quality monitoring, and apply cost controls such as lifecycle rules and storage classes to balance availability against expenditure.

Cloud StorageDataplexData lakeAccess control

Practice question for this objective

Free sampleStoring Datamedium

A genomics startup uses a Cloud Storage bucket as the landing zone of its data lake. Files arrive continuously and are read heavily for the first 30 days, occasionally between 30 and 180 days, and almost never after 180 days, but compliance requires the team to keep every object for at least seven years. The finance team wants to minimise storage cost without changing the application that writes and reads objects by their existing paths. Which configuration best meets these goals?

AManually copy objects older than 30 days into a separate Coldline bucket and update the application to read from both buckets depending on object age.
BEnable Autoclass on the bucket and additionally set a lifecycle rule that deletes objects after 180 days to control cost.
CConfigure an object lifecycle policy on the bucket that transitions objects to Nearline storage after 30 days and to Archive storage after 180 days, while leaving the bucket location and object names unchanged. Correct
DRecreate the bucket as Archive storage class from day one and rely on early-deletion fees being waived for compliance use cases to control short-term costs.

Apply Cloud Storage lifecycle rules to move data through storage classes that match its access pattern while preserving retention requirements. Cloud Storage lifecycle management evaluates configured conditions such as object age and transitions objects to colder storage classes in place, which preserves bucket name, object paths, and application behaviour. Combining a Nearline transition for cool data and an Archive transition for cold data aligns cost with access frequency while leaving the seven-year retention intact.

Why A is wrong: Manual copies double-handle data and force application changes to look up the right bucket, which is operationally fragile and does not exploit Archive pricing for the long tail of seven-year retention.

Why B is wrong: Autoclass would manage class transitions, but the additional deletion rule directly violates the seven-year retention requirement, so this configuration is not safe even though it appears cost optimal.

Why C is correct: Lifecycle rules transition objects between storage classes in place, so existing paths and the application remain unchanged while older objects move to cheaper Nearline and Archive classes that match their access frequency.

Why D is wrong: Archive class incurs higher retrieval and minimum-storage-duration charges for objects accessed in the first 30 days, and there is no general waiver of early-deletion fees, so heavily read recent objects would become expensive.

See more PDE practice questions, answers explained.

Exam traps in Storing Data

Answers that look right on this material and are not. Each one is a distractor from a different question in the PDE bank for this domain.

Manually rewrite each object to a colder storage class through a scheduled batch job that copies objects between a Standard bucket and a separate Coldline bucket once they pass an age threshold.
Why it is wrong: Copying objects between two buckets does move data to a colder class, but it changes the object paths that applications already use and adds a bespoke job to maintain, which breaks the requirement to leave the reading and writing applications untouched.
Apply uniform bucket-level access on every bucket and rely on legacy bucket ACLs to grant each persona the access pattern they require across the lake.
Why it is wrong: Uniform bucket-level access actually disables ACLs in favour of IAM, and ACLs cannot govern BigQuery datasets or Pub/Sub topics, so they cannot provide the single cross-service governance plane the security team needs.
A Dataform pipeline triggered by Cloud Scheduler that runs SQL assertions against the files through a BigQuery external table and emails the platform team when an assertion fails.
Why it is wrong: Tempting because Dataform supports assertions, but failures live inside Dataform runs and do not appear as catalog signals on the underlying Cloud Storage asset, which is the scenario's requirement.

Examworthy is not affiliated with or endorsed by Google Cloud. Original, blueprint-aligned practice material only.