This domain is about 18% of the exam. Questions are short business scenarios: a company wants to grow faster, cut cost or stop running data centers, and you pick the idea or benefit that fits. You are rarely asked about a specific product here. You are asked to use the vocabulary correctly and to match a stated goal to the right model.
Core definitions
| Term | Meaning |
|---|---|
| Cloud | Computing resources (compute, storage, networking, software) delivered over the internet on demand, with pay-for-what-you-use pricing and the provider running the underlying hardware |
| Infrastructure | The physical and virtual base that applications run on: servers, storage, network, data centers |
| Digital transformation | Using digital technology to change how a business operates and serves customers: products, processes, culture and customer experience, not just moving servers |
| Agentic AI | AI systems (agents) that can plan, use tools and take multi-step actions toward a goal with limited human direction, rather than only answering one prompt |
| Open source | Software whose source code is public and can be used, changed and shared under a licence (Kubernetes, TensorFlow, Android) |
| Open standard | A publicly documented specification that anyone can implement, so products from different vendors work together |
Digital transformation is not the same as digitization (turning paper into files) or a one-off lift-and-shift. It is a change in how the organization works, and technology is the enabler.
Benefits of cloud technology
Expect “which benefit does this scenario show?” questions.
| Benefit | What it means | Typical scenario |
|---|---|---|
| Scalability and elasticity | Add or remove capacity as demand changes | Retailer handles a holiday sales spike, then scales back |
| Cost-effectiveness | Pay for use; no big upfront hardware buy; less idle capacity | Startup avoids buying servers |
| Agility and speed | Provision in minutes, experiment cheaply, release faster | Team prototypes a new product without a procurement cycle |
| Flexibility | Choose services, regions and architectures; change your mind | Move from VMs to containers later |
| Enhanced security | Provider invests in layered security, patching and compliance at a scale most firms can’t match | Small company inherits Google-grade infrastructure protection |
| Global reach | Deploy near customers in many regions | Launch in a new country without building a data center |
| High availability | Redundant zones and regions keep services up | Service survives a data center failure |
| Data-driven insight | Analytics and AI services over all your data | Marketing uses real-time dashboards |
| Strategic value and focus | Staff stop racking servers and work on what differentiates the business | IT shifts from maintenance to innovation |
Elasticity is the ability to scale both up and down automatically. Scalability alone can mean only growing.
Drivers, challenges and the risk of standing still
What pushes organizations to transform
- Customer expectations for fast, personalized, always-on digital experiences.
- Competition from digital-native companies.
- Pressure to cut cost and free money from running hardware.
- Need to use data and AI for decisions.
- Business disruption (a pandemic, a supply chain shock) that demands agility.
- Regulation, security threats and ageing infrastructure.
Common hurdles
- Legacy systems that are hard to move or integrate.
- Skills gaps and resistance to change (culture is usually the hardest part).
- Unclear business goals or no executive sponsor.
- Data silos, security and compliance worries, and cost surprises.
- Vendor lock-in concerns.
Implications and risks of not adopting cloud
- Slower release cycles than competitors; lost market share.
- Higher cost of maintaining and refreshing hardware.
- Capacity that cannot absorb spikes, or that sits idle.
- Falling behind on AI and analytics.
- Growing security and compliance burden on an in-house team.
A good transformation starts with business goals and a clear owner, adds the right skills and change management, and moves in stages. Technology without culture change tends to stall.
Google Cloud differentiators
The guide names these. Learn each as one phrase.
| Differentiator | Meaning |
|---|---|
| World-leading AI | Google’s own research and first-party models (the Gemini family), available through Google Cloud and used in Google’s own products |
| Openness and interoperability | Built on open source and open standards (Kubernetes, Apache Beam, TensorFlow); services that run in hybrid and multicloud; fewer lock-in risks |
| AI Hypercomputer | An integrated supercomputing system for AI: purpose-built hardware (GPUs and TPUs), software, and flexible consumption models |
| AI-ready data platform | Data and analytics services (BigQuery and others) that connect directly to AI so data does not have to be moved around |
| Security | Secure-by-design infrastructure, defense in depth, and Google threat intelligence |
| Global network | One of the largest private fiber networks, carrying Google’s own traffic and customers’ traffic with low latency |
The word to watch for is open. When a scenario worries about being trapped with one vendor, the Google Cloud answer is openness, open source and multicloud support.
Private cloud, hybrid cloud and multicloud
| Model | What it is | Why choose it |
|---|---|---|
| Public cloud | Resources owned and run by a provider, shared among customers (logically isolated) | Lowest upfront cost, fastest scale |
| Private cloud | Cloud-style infrastructure used by one organization, on premises or hosted | Strict control, data residency, regulatory demands, existing investment |
| Hybrid cloud | A mix of on-premises or private infrastructure with public cloud, connected so workloads can span both | Keep sensitive or legacy systems on premises while using cloud for new workloads; migrate gradually; burst to the cloud at peak |
| Multicloud | Using more than one public cloud provider | Avoid lock-in, use each provider’s best service, resilience, acquisitions that brought another cloud |
Hybrid and multicloud are different. Hybrid = on premises plus cloud. Multicloud = more than one public cloud. An organization can be both. Google Cloud supports both through products built to run anywhere (see the modernizing infrastructure notes).
Trap: private cloud is not automatically more secure. Security depends on how it is run. Pick private cloud when the scenario stresses control or regulation, not because “private sounds safer”.
Networking basics
| Term | Meaning |
|---|---|
| IP address | A numeric label identifying a device or service on a network. IPv4 looks like 203.0.113.10; IPv6 is longer. Public addresses are reachable from the internet, private ones only inside a network |
| DNS (Domain Name System) | The internet’s phone book: turns names such as example.com into IP addresses. Google Cloud’s managed service is Cloud DNS |
| Latency | The time a request takes to travel and be answered. Lower is better. Users near a service see lower latency |
| Bandwidth | How much data can be moved per second. A “wide pipe” |
| Throughput | The data actually moved per second (bandwidth is the ceiling) |
Latency vs bandwidth is the classic confusion. Lots of bandwidth does not cut the delay of one round trip. To reduce latency, put the service closer to users (regions, edge caching).
Google Cloud’s network infrastructure
Google runs a global private network of fiber, undersea cables and points of presence. Traffic between Google Cloud locations travels on Google’s own network rather than the public internet, which means more consistent performance, lower latency and better security.
Locations, from largest to smallest
| Component | What it is |
|---|---|
| Region | An independent geographic area, such as us-central1 or europe-west1, with multiple zones. Choose a region for latency to users, data residency rules, price and which services are available |
| Zone | A deployment area inside a region (a data center or group of them), such as us-central1-a. Zones in a region are isolated from each other’s failures. Run in two or more zones for high availability |
| Multi-region | A large area spanning several regions (for example “US” or “EU”) used by some services such as Cloud Storage and BigQuery for higher availability |
| Edge location (point of presence) | A location near users where traffic enters Google’s network and content can be cached by Cloud CDN, which lowers latency |
How they work together:
- Spread across zones to survive a zone failure.
- Spread across regions to survive a regional outage, or to serve distant customers.
- Use edge locations and a CDN to serve static content close to users.
- Keep data in a chosen region for data residency.
Trap: a zone is not a region. A multi-zone design is not protected against a whole-region failure.
Service models: IaaS, PaaS, SaaS
The models differ by how much you manage versus how much the provider manages. This is called the shared responsibility split.
| Model | You manage | Provider manages | Google Cloud example | Benefit | Trade-off |
|---|---|---|---|---|---|
| IaaS | OS, runtime, apps, data | Hardware, network, virtualization | Compute Engine | Most control and flexibility; easiest lift and shift | Most operational work |
| PaaS | Your application and data | Hardware, OS, runtime, scaling | App Engine, Cloud Run, GKE (managed Kubernetes) | Developers focus on code; faster delivery | Less control over the platform |
| SaaS | Just using it (and your data and settings) | Everything | Google Workspace, Looker, Gemini Enterprise | Fastest to adopt, no maintenance | Least customization |
A quick way to read the scenario:
- “We want full control of the operating system” -> IaaS.
- “We want developers to deploy code and not worry about servers” -> PaaS.
- “We want an off-the-shelf application with no infrastructure” -> SaaS.
As you move from IaaS to SaaS, control goes down and provider responsibility goes up. Customers always stay responsible for their data, identities and access in every model.
Exam checklist
- Define digital transformation as a business change enabled by technology, not just migration.
- Match scenarios to cloud benefits: spikes = elasticity, no upfront buy = cost-effectiveness, fast experiments = agility, new country = global reach.
- Know the Google Cloud differentiators: AI, openness, AI Hypercomputer, AI-ready data, security, global network.
- Hybrid is on premises plus cloud. Multicloud is more than one public cloud. Private cloud is single-organization.
- Latency is delay; bandwidth is capacity. Cloud DNS is the DNS service.
- Region contains zones; spread across zones for high availability, across regions for disaster recovery; edge locations cut latency.
- IaaS = most control, PaaS = focus on code, SaaS = ready-made app.
- Customers always own their data and access decisions.
- Cultural change and skills are usually the hardest transformation hurdles.