Complete guide
CDL Cloud Digital Leader
Tap any service to see more details and official docs.
Exam breakdown
40–60
Questions
~50
Min read
700
Passing score
Digital transformation with Google Cloud
17% of the examWhat is digital transformation?
Digital transformation is the process of using technology to improve performance, reach new goals or create new business models.
Reimagine
Redesign business processes leveraging cloud capabilities that don't exist on-premises.
Innovate
Launch new products and services faster thanks to the agility of the cloud.
Scale
Grow into global markets without investing in physical infrastructure in every region.
Why Google Cloud
| Strength | Detail |
|---|---|
| Global network | Private submarine fiber-optic network connecting every datacenter. Low latency worldwide. |
| Security by design | Infrastructure designed from the ground up for security. Titan chip, BeyondCorp Zero Trust. |
| AI & ML | Leaders in AI: TensorFlow, TPUs, Vertex AI, Gemini. Years of lead in foundation models. |
| Sustainability | Carbon neutral since 2007. Goal of 24/7 carbon-free energy by 2030. |
| Data & analytics | BigQuery processes petabytes in seconds. Looker for BI. Leaders in analytics at scale. |
| Open source | Kubernetes, TensorFlow, Istio — all created by Google and donated to the community. |
Google Cloud global infrastructure
| Component | Description |
|---|---|
| Regions | Geographic areas with multiple zones. More than 40 regions. Choose by latency, compliance or price. |
| Zones | 3+ zones per region, physically separated. High availability by deploying across multiple zones. |
| Edge Network | More than 100 points of presence for Cloud CDN and Cloud Interconnect. Minimal latency to the end user. |
| Private network | All communication between GCP services travels over Google's private network, never the public internet. |
| Multi-region | US, EU, APAC: groups of regions for automatic data replication with maximum availability. |
A region has multiple zones (at least 3). For high availability deploy across multiple zones. For disaster recovery, across multiple regions. Multi-region locations are only for Cloud Storage and a few managed services.
Exploring data with Google Cloud
16% of the examStorage and databases
Click any service to see more details
| Service | Type | Use case |
|---|---|---|
| Cloud Spanner | Global SQL | Globally distributed relational database. ACID transactions at massive scale. |
| Bigtable | NoSQL columnar | Petabytes of time-series data, IoT, AdTech. Sub-10ms low latency. |
| Memorystore | In-memory cache | Managed Redis and Memcached. For sessions, query caching and simple pub/sub. |
| AlloyDB | PostgreSQL+ | PostgreSQL-compatible, 4x faster for analytical workloads. Premium alternative to Cloud SQL. |
BigQuery: analytics at scale
Click any service to see more details
| Feature | Detail |
|---|---|
| Architecture | Columnar storage separated from compute. Both can scale independently. |
| BigQuery ML | Train and deploy ML models with SQL. No moving data to another platform. |
| Looker Studio | Free BI tool that connects directly to BigQuery for dashboards. |
| BigQuery Omni | Query data in AWS S3 or Azure Blob Storage from BigQuery using SQL. |
| Streaming | Insert data in real time with second-level latency for near-live analysis. |
Pipelines and data processing
| Service | Function |
|---|---|
| Dataflow | Batch and streaming data processing. Based on Apache Beam. Serverless. |
| Dataproc | Managed Hadoop and Spark clusters. For existing BigData workloads. |
| Cloud Data Fusion | Visual ETL with a drag-and-drop interface. No code for data integrations. |
| Looker | Enterprise BI platform with LookML. For business data analysis and exploration. |
| Dataplex | Unified management of distributed data. Governance, quality and data discovery. |
The standard streaming pattern on GCP is: data source → Pub/Sub → Dataflow → BigQuery. For batch: Cloud Storage → Dataflow → BigQuery. BigQuery can also read directly from Cloud Storage.
AI and innovation with Google Cloud
16% of the examPre-trained AI APIs
| API | Function | When |
|---|---|---|
| Vision API | Detects objects, faces, text (OCR), content moderation in images | Analyzing images without training your own model |
| Natural Language API | Sentiment analysis, entities, syntax and text classification | NLP without custom ML: reviews, support tickets |
| Speech-to-Text | Transcribes audio to text in 125+ languages with automatic punctuation | Transcribing calls, meetings, voice commands |
| Text-to-Speech | Natural voice synthesis with WaveNet and Studio voices | Voice assistants, content read-aloud |
| Translation API | Automatic text translation in 100+ languages with automatic detection | App localization, multilingual support |
| Video Intelligence API | Detects objects, actions and text within videos in real time | Content moderation, video analysis |
| Document AI | Extracts structured data from documents (invoices, contracts, forms) | Automating document-based processes |
Vertex AI: ML platform
Click any service to see more details
AutoML
No-code models: tabular, images, text, video. Google searches for the best architecture automatically.
Custom Training
Train with your own code (TensorFlow, PyTorch, scikit-learn) on Google-managed hardware.
Vertex Workbench
Managed Jupyter notebooks with access to GPUs and TPUs for data exploration and experiments.
Model Registry
Versioning and lifecycle management for models: registration, monitoring and endpoint deployment.
Generative AI with Google Cloud
| Service/Concept | Description |
|---|---|
| Gemini | Google's family of multimodal models (text, images, code, video, audio). Available on Vertex AI. |
| Vertex AI Studio | Interface to test and prompt-engineer with Gemini models and other foundation models. |
| Model Garden | Catalog of foundation models from Google and third parties (Llama, Mistral, etc.) ready to use. |
| Gemini for Google Workspace | Generative AI built into Gmail, Docs, Sheets, Meet. For enterprise productivity. |
| Grounding | Ground AI responses with your own data to reduce hallucinations. Connects to Cloud Search or data in GCS. |
| Embeddings API | Generates numerical text vectors for semantic search, recommendations and RAG. |
For the CDL: distinguish between pre-trained APIs (you use Google\'s model as-is), Vertex AI AutoML (train with your data, no code), and Vertex AI Custom Training (your own code). For GenAI: Gemini is Google Cloud\'s main foundation model.
Modernizing infrastructure and applications
17% of the examCompute on Google Cloud
Click any service to see more details
| Service | Abstraction | When |
|---|---|---|
| App Engine | PaaS | Web apps with no infrastructure to manage at all. Standard (stateless) or Flexible (containers). |
| Bare Metal | Hardware | Dedicated physical servers for specialized workloads like Oracle databases. |
| VMware Engine | VMware as a Service | Migrate on-premises VMware workloads without refactoring. Keeps existing VMware tooling. |
Networking on Google Cloud
| Service | Function |
|---|---|
| VPC | GLOBAL private virtual network (the key difference from AWS). A VPC can have subnets in every region. |
| Cloud Load Balancing | Global HTTP/S load balancing with Anycast IP. A single VIP that routes to the nearest backend in the world. |
| Cloud CDN | Content delivery network integrated with Cloud Load Balancing. Caches at Edge Locations. |
| Cloud DNS | Highly available managed DNS with 100% SLA. Integrated with Cloud Load Balancing. |
| Cloud Interconnect | High-speed private network connection between on-premises and GCP. Alternative to VPN for higher bandwidth. |
| Cloud VPN | Encrypted IPsec connection between on-premises and VPC. For workloads with lower bandwidth requirements. |
Migration and modernization
| Strategy | Description | Effort |
|---|---|---|
| Rehost (Lift & Shift) | Move on-premises VMs to Compute Engine with no changes. Fast but doesn't leverage cloud-native features. | Low |
| Replatform (Lift & Reshape) | Small changes to use managed services (e.g. migrating MySQL to Cloud SQL). | Medium |
| Refactor | Refactor code to take advantage of cloud features (containers, microservices, PaaS). | High |
| Rearchitect | Redesign the application from scratch with cloud-native architecture (serverless, event-driven). | Very high |
| Replace | Replace the application with an equivalent SaaS (e.g. migrating Exchange to Google Workspace). | Variable |
Key migration tools
Migrate to VMs
Migrates VMware/AWS/Azure VMs to Compute Engine
Database Migration
Migrates MySQL/PostgreSQL to Cloud SQL with minimal downtime
BigQuery Migration
Migrates Teradata, Redshift data warehouses to BigQuery
GKE multi-cloud / GDC
GKE (fleets, GKE Multi-Cloud) and Google Distributed Cloud manage apps on GCP, on-premises and other clouds (formerly Anthos / GKE Enterprise)
All six domains weigh between 16% and 17%. Memorize the difference between Compute Engine (VMs/IaaS), GKE (Kubernetes), Cloud Run (serverless containers) and Cloud Run functions (serverless functions). The exam will give a scenario and ask which to choose.
Trust and security with Google Cloud
17% of the examShared responsibility model
🏢 Google — Security OF the cloud
- • Physical hardware and datacenters
- • Global private network
- • Hypervisor and platform
- • Managed services software
- • Titan chip for boot integrity
👤 Customer — Security IN the cloud
- • Customer data
- • IAM, users and access
- • Application configuration
- • Compute Engine OS (IaaS)
- • Customer data encryption
Note: on PaaS (Cloud Run, App Engine) and SaaS (Google Workspace), Google takes on more responsibility. On IaaS (Compute Engine) the customer takes on more.
IAM and access control
Click any service to see more details
| Concept | Description |
|---|---|
| Principal | Who: Google account, group, Service Account, Workspace domain, federated identity |
| Role | What it can do: a collection of permissions. Basic / Predefined / Custom |
| Policy | Binds principals to roles on a specific resource |
| Service Account | Identity for apps and VMs, not for people. Accesses APIs without hardcoded credentials |
| Workload Identity | Lets GKE pods act as Service Accounts without managing keys |
| IAM Conditions | Conditional access by time, IP, resource tag, etc. |
Security services
| Service | Function |
|---|---|
| Cloud Armor | WAF and DDoS protection for resources behind Cloud Load Balancing. |
| Security Command Center | Centralized visibility into threats, vulnerabilities and compliance across the GCP organization. |
| Cloud KMS | Key Management Service: manages encryption keys. CMEK to encrypt data with your own keys. |
| Secret Manager | Stores and manages secrets (API keys, passwords, certificates) with automatic rotation. |
| Google Security Operations | Google's SIEM and SOAR for threat detection at scale. Analyzes security logs with AI. |
| Chrome Enterprise Premium | Zero Trust Network Access. Verifies identity and device, not network location. |
| VPC Service Controls | Security perimeters for sensitive data. Prevents data exfiltration outside the perimeter. |
Shared Responsibility: on Compute Engine (IaaS), the customer is responsible for the operating system, patches and network configuration inside the VM. On Cloud Run or App Engine (PaaS), Google manages the OS. On Google Workspace (SaaS), Google manages the entire stack except the customer's data.
Scaling with Google Cloud operations
17% of the examGoogle Cloud Observability
Click any service to see more details
| Service | Function |
|---|---|
| Cloud Monitoring | Performance metrics, custom dashboards, alerts and SLO tracking. |
| Cloud Logging | Stores and analyzes logs from every GCP service. Cloud Audit Logs for auditing. |
| Cloud Trace | Distributed latency analysis. Finds bottlenecks in microservices. |
| Cloud Profiler | Analyzes CPU and memory consumption of production code with minimal overhead. |
| Error Reporting | Aggregates and prioritizes application errors automatically. Notifies about new errors. |
Cost management on Google Cloud
Cloud Billing
Centralized billing management. Configure budget alerts and export cost data.
Cost Management Tools
Console tools to see costs broken down by service, project or label.
Budget Alerts
Set budgets and receive alerts when spending exceeds defined thresholds.
Recommender
AI that suggests cost optimizations: underutilized VMs, idle resources, applicable CUDs.
| Discount | Description | Savings |
|---|---|---|
| Sustained Use Discounts (SUD) | Automatic: for using a VM >25% of the month. No commitment. | Up to 30% |
| Committed Use Discounts (CUD) | 1-3 year commitment for specific resources. | Up to 55% (70% on memory-optimized) |
| Spot VMs (formerly preemptible) | Interruptible VMs from Google's spare capacity. | Up to 91% |
| Free Tier | Monthly free usage: Cloud Storage, Compute Engine, BigQuery, etc. | Variable |
SRE and reliability principles
| Concept | Definition |
|---|---|
| SLI (Service Level Indicator) | A specific metric that measures the health of the service (e.g. error rate, p99 latency) |
| SLO (Service Level Objective) | The target for the SLI (e.g. "99.9% of requests under 200ms over the last 30 days") |
| SLA (Service Level Agreement) | Contract with the customer. If the SLO isn't met, there are financial consequences. |
| Error Budget | Amount of allowed failure = 100% - SLO. Balances development speed vs reliability. |
| MTTR / MTBF | MTTR = mean time to recovery. MTBF = mean time between failures. |
SUD (Sustained Use Discounts) are automatic — Google applies the discount with no action on your part. CUD requires an explicit commitment. Spot VMs are cheap but your app must tolerate interruptions.
Tips for exam day
IaaS vs PaaS vs SaaS on GCP: Compute Engine=IaaS, App Engine/Cloud Run=PaaS, Google Workspace=SaaS. The exam gives scenarios to pick the right model.
Resource hierarchy: Organization → Folder → Project → Resource. IAM policies are inherited downward. A project is the base unit of billing and isolation.
Pre-trained AI APIs: Vision API=images, Natural Language API=text, Speech-to-Text/Text-to-Speech, Translation API, Video Intelligence API. No custom ML required.
Right storage choice: Firestore=NoSQL docs, Cloud SQL=relational SQL, Bigtable=NoSQL columnar at scale, Spanner=global SQL, Cloud Storage=objects, Memorystore=cache.
Security: IAM controls who can do what. Cloud Armor protects HTTP/S. VPC Service Controls adds security perimeters. Security Command Center centralizes visibility.
BigQuery is serverless. Billed by data processed in queries. Use LIMIT and specific SELECTs to reduce costs. For real-time dashboards consider BI Engine.
Compute: Compute Engine=VMs, GKE=Kubernetes, Cloud Run=serverless containers, Cloud Run functions=serverless functions, App Engine=web apps with no infrastructure to manage.
Networking: VPC is global (not per-region like AWS). A single VPC can have subnets in multiple regions. Cloud Load Balancing is global and anycast.
Billing per minute (not per hour). Automatic Sustained Use Discounts (up to 30%). Committed Use Discounts (up to 55%, 70% on memory-optimized). Spot VMs up to 91% discount.
Modernization: the 6 Rs of migration (Rehost, Replatform, Refactor, Rearchitect, Rebuild, Replace). GKE (fleets, GKE Multi-Cloud) and Google Distributed Cloud manage workloads across GCP, on-premises and other clouds.
SRE and operations: SLI=health metric, SLO=target for the SLI, SLA=contract with a customer. Error budget = 100% - SLO. Cloud Monitoring and Logging are part of Google Cloud Observability.
Shared responsibility: Google manages the physical infrastructure, network and managed services. The customer manages data, IAM access, app configuration and data encryption.
Want to go deeper?
This guide is a quick summary. Visit the Deep Dive section for an exhaustive walkthrough with real cases, cost analysis and architecture decisions.
Practice CDL with interactive simulations
Click-through labs: navigate a step-by-step portal simulation, with clear objectives and verification on every action — no real costs.
Ready to practice?
Put what you learned to the test with original CDL exam questions.
Start CDL mock examUnlimited mock exams with the Pro plan
Was this guide helpful?
Not ready to start yet?
Leave your email and create your free account whenever you're ready — start with 15 CDL questions, no credit card required.
Already have an account? Go straight to the mock exam →