2026.06.23 2026.06.23

Cloud Digital Leader Part 1: Exam Overview and the First Three Domains

swiftwand

Cloud Digital Leader (CDL) is Google Cloud’s only Foundational credential, and it rewards business-level cloud literacy rather than hands-on implementation. This first part covers the exam’s overall shape and the front three of its six domains — digital transformation, data, and AI/ML — which together form the scoring base you build the rest of the exam on.

The shape of the Cloud Digital Leader exam — six domains and weights
Domain 1 — digital transformation and cloud fundamentals (~17%)
Domain 2 — turning data into value (~16%)
Domain 3 — remember AI/ML as three choices (~16%)
Using your AWS knowledge to clear the front three domains fast
Where AWS veterans stumble on the front three domains
Conclusion — lock in the scoring base with the front three domains
References

忍者AdMax

The shape of the Cloud Digital Leader exam — six domains and weights

Item	Spec
Level	Foundational
Duration	90 minutes
Questions	50-60 (multiple choice)
Fee	USD 99 (plus tax; about 15,800 yen at ~160 yen/USD)
Validity	3 years
Delivery	Online-proctored / test center
Prerequisites	None

The exam spans six domains. The first three — covered here — carry roughly 17%, 16%, and 16% of the score. Get these right and you have secured close to half the exam before touching the infrastructure-heavy back half.

Domain 1 — digital transformation and cloud fundamentals (~17%)

This domain frames why organizations move to the cloud and what the basic service models mean. The single most testable idea is the split between IaaS, PaaS, and SaaS, and the shared responsibility model that goes with each.

Model	What you manage	Mental model	Operating burden
IaaS	You rent infrastructure such as VMs and storage	You manage the OS and everything above it	High freedom, high burden
PaaS	The runtime is provided for you	Focus on the app and its data	Moderate
SaaS	You use finished software	Configure and use only	Lightest

Domain 2 — turning data into value (~16%)

Google Cloud’s identity is built on data, so this domain asks you to pick the right managed data product for a given workload. Learn this short cheat sheet and most data questions resolve to a single best answer.

Product	Type	Primary use
Cloud Storage	Object storage	Unstructured data, backup, delivery
Cloud SQL	Managed relational DB	Small-to-mid business systems
Cloud Spanner	Globally distributed relational DB	Large, global, strongly consistent workloads
Cloud Bigtable	Wide-column NoSQL	High-volume time-series, IoT, analytics
Firestore	Document NoSQL	Mobile and web app data
BigQuery	Serverless data warehouse	Large-scale analytics, BI, SQL analysis

Domain 3 — remember AI/ML as three choices (~16%)

The AI/ML domain is easiest to hold in your head as a ladder of three options by effort and customization: pre-trained APIs (Vision, Translation, and similar) for ready-made capabilities, low-code tools for custom models without deep ML skills, and the full AI platform — Vertex AI, now part of the Gemini Enterprise Agent Platform — for teams that build and train their own models. Match the business need to the lowest-effort option that satisfies it, and you answer most questions correctly.

Using your AWS knowledge to clear the front three domains fast

If you hold an AWS credential, most of this material is a renaming exercise rather than new learning.

AWS knowledge	CDL equivalent
Shared responsibility model	Tested with the same logic
S3 storage classes	Cloud Storage Standard / Nearline / Coldline / Archive
Redshift	BigQuery (more serverless in orientation)
RDS	Cloud SQL
DynamoDB	Firestore / Bigtable
SageMaker	Gemini Enterprise Agent Platform (formerly Vertex AI)
Pre-trained AI services	Pre-trained API set (Vision, Translation, and more)

Where AWS veterans stumble on the front three domains

Three traps recur. First, BigQuery is positioned more aggressively as serverless than Redshift, so questions reward you for choosing it when the scenario stresses zero infrastructure management. Second, Google splits NoSQL across Firestore and Bigtable, and the exam expects you to pick by access pattern — document versus wide-column — rather than treating them as one DynamoDB-shaped box. Third, the AI/ML domain wants the lowest-effort option that meets the requirement, so reaching for the full platform when a pre-trained API would do is a common wrong answer.

Conclusion — lock in the scoring base with the front three domains

Domains 1 through 3 reward conceptual clarity over memorization: the service models, the data-product cheat sheet, and the three-tier AI/ML ladder. Anchor those, map them to the AWS terms you already know, and you carry roughly half the exam into the back three domains covered in Part 2.