Prepare and test your skills

Prepare and test your skills

Cloud transformation is the shift from owning and managing physical servers to using on-demand automation, managed services, and global infrastructure. Traffic flows from users to Google Cloud services that add or remove capacity automatically, while spending shifts from upfront hardware purchases to usage-based payment. This change affects daily operations, team skills, and budgets, and it opens new work with data analytics, machine learning, and artificial intelligence (AI).
The four phases of cloud migration are the ordered steps that move workloads from review to ongoing improvement. First, Assess inventories applications, dependencies, costs, and performance benchmarks. Then, Plan builds the foundation with identity management, organization and project structure, and networking, plus a prioritized migration strategy. Next, Deploy moves workloads and data to Google Cloud. Finally, Optimize improves performance, scalability, disaster recovery, and cost and opens the door to AI. A weak Plan phase can cause delays, confusion, and downtime because later steps depend on that foundation.
Elastic operations replace manual, fixed-capacity work with services that scale based on policy. Traffic flows from users to the scaling resource, which then adds or removes capacity:
Right-sizing keeps resources matched to demand. Cloud Monitoring collects real-time metrics, logs, and traces, Recommender suggests better sizes, and Cloud Build and Cloud Deploy create CI/CD pipelines where automated tests validate changes before production.
Cloud-native design builds applications from small, independent parts that scale and recover quickly. A decoupled architecture lets each microservice upgrade and scale on its own, while a stateless design stores data in shared storage instead of local disks so new instances start fast and survive restarts. Autoscaling policies watch signals such as CPU, memory, and queue depth to add or remove resources. Data flows from services into Cloud Monitoring for real-time insight, then into BigQuery for storing and analyzing trends that guide tuning such as caching, indexing, and code fixes.
Open approaches reduce dependence on one vendor but solve different problems. Open source software shares code the community can use and improve, such as Kubernetes, while open standards share public rules many vendors follow so products work together. A hybrid architecture connects on-premises systems to Google Cloud with VPN or Interconnect, so legacy data can support new cloud apps through APIs managed by Apigee. Democratized data extends the same idea to people, where data flows from many sources into tables in BigQuery, Cloud Data Fusion (opens in a new tab) builds code-free pipelines to cleanse and prepare data, and Looker provides governed dashboards from the same trusted version.
Exam tip: Open source software is code you can inspect and modify; open standards are rules that different products follow so they can work together.
Business value from Google Cloud groups into scalability to grow, flexibility to run anywhere, agility to release fast, security to lower risk, cost efficiency to avoid waste, and strategic value to compete faster. Scalability lets Cloud Run, Compute Engine, and Firestore grow during peaks and shrink when quiet. Flexibility lets workloads run on premises, in private hosting, in public cloud, or in hybrid and multicloud combinations. Intelligence turns large datasets into actions with BigQuery, Looker, and location tools, while collaboration connects people through Google Workspace with Cloud Identity and Identity-Aware Proxy (IAP) for secure access.
Placement choices trade control for speed and resilience. On-premises infrastructure keeps control in a private data center for legacy or regulated workloads, public cloud adds quick setup and elasticity, private cloud dedicates resources to one organization, hybrid links private and public so requests and data flow both ways, and multicloud uses more than one provider for sovereignty, unique services, and outage resilience. Drivers for change include market speed, customer quality, analytics and AI, and balanced cost, while barriers include app dependencies, resistance to DevOps and Site Reliability Engineering (SRE), and skill gaps. Without modernization, releases stay slow and manual, customer experience suffers without feedback and real-time insight, and outdated systems cost more because monoliths must be bought for peaks yet sit idle otherwise.
Continue with this study guide: Why Cloud Technology is Transforming Business
General cloud concepts explain how spending, scaling, placement, and responsibility shape cost and control. The economic shift moves spending from owned hardware to usage-based billing, while service models and deployment choices define who manages what and where workloads run. Google Cloud supports this shift with elastic capacity, shared pools of resources, and options from zonal to global availability.
The shift from capital expenditure (CapEx) to operational expenditure (OpEx) changes when and how IT is paid for. In CapEx, the organization buys servers, storage, licenses, power, cooling, and rack space before applications run and must buy for peak demand, so much hardware sits idle. In OpEx, the organization pays for consumed VMs, containers, and storage and releases them when demand drops. Total cost of ownership (TCO) is the complete lifetime cost, including hardware, power, patching, scaling, security, disaster recovery, migration, network transfer, and support.
| Dimension | CapEx on premises | OpEx in cloud |
|---|---|---|
| Payment timing | Upfront purchase before use | Ongoing charges based on measured use |
| Capacity | Fixed, bought for peaks | Variable, adjusted by autoscaling and right-sizing |
| TCO inputs | Hardware, facilities, salaries | Compute, Storage, Network, Operation, licenses, migration, and discounts |
Long-term control follows a fixed order: first monitor costs with billing reports, labels, and budget alerts; then control provisioning with policies and ownership; then reduce costs with discounts, right-sizing, autoscaling, and data lifecycle management.
Scalability is the ability to handle more or less work by adding or removing resources, elasticity is automation of that change in response to demand, and reliability is continuing to function despite failures. Traffic flows from users to the autoscaling control and then to the added capacity, and the reverse happens when demand falls. Compute Engine MIGs automate VM changes within minimum and maximum limits, GKE scales nodes and Pods together, and Cloud Run scales serverless containers directly with incoming traffic. On-demand self-service lets teams provision from shared pools of compute, storage, and networking without waiting for manual setup.
Cost optimization lowers the OpEx rate while keeping flexibility:
Exam tip: Choose Spot VMs when the workload is fault-tolerant and can be interrupted; choose CUDs when usage is steady and predictable over one or three years.
Deployment models describe where workloads run and how placement affects availability, cost, and complexity. Public cloud runs entirely in one provider, private cloud runs in a dedicated environment, hybrid combines on-premises with Google Cloud, and multicloud combines Google Cloud with other providers. Hybrid and multicloud add networking, monitoring, and security work because tools and controls are not standard across environments. A common sovereignty case forces multicloud: if the current provider has no local region in a country, the compliant solution is to use another provider that has one.
Google Cloud availability targets trade simplicity and cost for resilience with synchronous replication:
| Archetype | Availability target | Outage resilience |
|---|---|---|
| Zonal | 99.9% | Low cost and simple, but recovery can take hours or days |
| Regional | 99.99% | Survives zone outages with near-zero recovery time objective (RTO) |
| Multi-regional and global | 99.999% | Survives zone and region outages with near-zero RTO |
The shared responsibility model divides work between provider and customer. The provider secures cloud infrastructure and supplies high availability, backups, and scalability backed by service level agreements (SLAs). The customer secures data, configurations, access, and cross-platform controls placed in the cloud.
The global backbone is Google Cloud's worldwide private network that carries traffic between regions and enables global load balancing. The backbone is built from owned fiber, public fiber, and subsea cables that move high volumes of data between continents. Points of presence (PoPs) are network edge locations where customer traffic enters or leaves the Google network. An internet service provider (ISP) gives users initial internet access, then traffic flows from the user device to the nearby PoP and across the private backbone to the destination region, which reduces public hops, lowers delay, and limits interception.
Addressing and geography organize that journey. An IP address is the routable numeric destination for packets, while Domain Name System (DNS) translates human-friendly names into those addresses so traffic flows from the requester to the correct resource. Cloud DNS serves public zones visible to the internet and private zones visible only inside a network. A region is a geographic location for resources, a zone is an isolated area inside a region, and edge locations in more than 200 metropolitan areas bring traffic onto Google's network early. Latency is round-trip delay, roughly 1 millisecond per 100 kilometers, while bandwidth is how much data can move. Ingress into Google Cloud is free for most services, egress leaving Google Cloud incurs charges for services such as Cloud Storage (opens in a new tab), Compute Engine, and Cloud SQL, and cross-zone or cross-region traffic can also incur charges. Regional Persistent Disk volumes replicate synchronously across two zones, while dual-region buckets, multi-region Spanner, and snapshots to multi-regional storage survive extended loss of an entire region.
Continue with this study guide: Fundamental Cloud Concepts
Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) form a spectrum from most customer control to least operational work. As movement goes from IaaS to PaaS to SaaS, the provider manages more and the customer does less. More control means more setup and upkeep, while more abstraction means faster deployment with less flexibility.
Infrastructure as a Service (IaaS) gives the customer the most control over virtualized resources. In IaaS, the customer manages the operating system, middleware, and applications, while Google manages virtualization, servers, storage, and networking. Compute Engine creates VM and bare metal instances with choices of machine series, storage, accelerators such as GPUs or TPUs, and guest operating systems. Google Cloud VMware Engine runs VMware workloads natively without re-architecting. The tradeoff is high operational overhead for patching, configuring, securing, and scaling.
Managed platforms that form PaaS let customers ship code while Google runs the platform. In PaaS, Google manages infrastructure, operating system, and runtime so customers focus on code, containers, and jobs. Work flows from customer code or containers to Google-managed infrastructure, which provisions resources and manages the lifecycle. Peer options include Batch for large batch jobs, Cloud Run for containers that respond to web requests or events, Cluster Director for AI and high performance computing clusters, and Google Kubernetes Engine (GKE) where Google manages the control plane. GKE Standard gives more infrastructure control, GKE Autopilot abstracts more, and Cloud Run abstracts away all infrastructure. SaaS delivers a complete application where the provider manages everything and the customer only configures settings and supplies data, such as Gmail, Google Workspace, Terraform Cloud, and Cloud Identity for single sign-on.
| Model | Customer manages | Provider manages | Example services |
|---|---|---|---|
| IaaS | Operating system, middleware, applications | Virtualization, servers, storage, networking | Compute Engine, VMware Engine |
| PaaS | Workloads and application code | Infrastructure, operating system, runtime | Batch, Cloud Run, GKE |
| SaaS | Use of the application and data inside it | Everything, including the application | Google Workspace, Cloud Identity |
Exam tip: In IaaS you manage the operating system and up, in PaaS the provider manages the operating system and runtime so you manage only workloads, and in SaaS the provider manages everything and you just use the application.
Choosing a model means matching control needs, team skills, delivery speed, and technical fit. First, identify control needs for operating system or server settings; if deep control is required, IaaS is the likely fit. Second, assess whether staff can manage those layers; if not, move toward PaaS or SaaS. Third, weigh delivery speed, since PaaS and SaaS remove setup work. Fourth, confirm technical fit, because PaaS and SaaS limit changes to the underlying environment.
Migration approaches align with those models. Rehost moves workloads with minimal changes to cloud VMs, replatform makes limited optimizations such as moving to Cloud SQL, refactor improves workloads for scalability, re-architect changes code such as turning a monolith into microservices, rebuild rewrites the app, and repurchase replaces it with SaaS such as Google Workspace. Lift-and-shift of an existing server fits IaaS because it preserves behavior, new cloud-native development fits PaaS because developers deploy without provisioning servers, and standard email or collaboration fits SaaS because users consume it directly. TCO comparison should use cost per unit of work, such as requests processed, because personnel and overhead are part of cost. Traffic flows from users through Cloud Load Balancing (opens in a new tab) to the frontend, which then reads and writes data in Cloud SQL.
The shared responsibility model divides security and management work between Google Cloud and the customer. Google Cloud protects data centers, hardware, networking, and global infrastructure, applies default encryption to data, and enforces the Identity and Access Management (IAM) policies that customers define. Enforcement flows from Google Cloud to customer workloads, so only approved identities can administer workloads or access data. The customer is always responsible for securing data, controlling who can access it, and correctly configuring applications and policies.
| Model | Google Cloud manages | Customer manages |
|---|---|---|
| On premises | Nothing | Everything, including physical security and data |
| IaaS | Physical security and underlying infrastructure | Operating systems, applications, data, IAM, and network settings |
| PaaS | Infrastructure, operating system, runtime, plus shared controls | Data security, client protection, and shared app and IAM configuration |
| SaaS | Infrastructure, platform, and application | Access controls and data stored in the application |
Customer duties never shift for data protection, least privilege, and workload configuration. The customer decides where data may reside, classifies sensitive data such as personally identifiable information (PII), grants only needed permissions with dedicated service accounts, and secures source code, images, and containers. In GKE Standard, Google Cloud manages the operating system underneath containers, but the operating system inside customer containers and the application code remain customer responsibilities. Governance adds organization policies, Assured Workloads for regulated needs, and continual monitoring with audit logs and posture tools.
Continue with this study guide: Cloud Computing Models and Shared Responsibility
Capital expenditure (CapEx) means buying servers, storage, licenses, power, cooling, and rack space upfront before applications run and buying for peak demand so much hardware sits idle, while operational expenditure (OpEx) means paying for consumed virtual machines, containers, and storage and releasing them when demand drops. Total cost of ownership (TCO) is the complete lifetime cost including hardware, power, patching, scaling, security, disaster recovery, migration, network transfer, and support.
The four phases in order are Assess, Plan, Deploy, and Optimize. Assess inventories applications, dependencies, costs, and performance benchmarks, Plan builds the foundation with identity management, organization and project structure, and networking plus a prioritized migration strategy, Deploy moves workloads and data to Google Cloud, and Optimize improves performance, scalability, disaster recovery, and cost and opens the door to artificial intelligence (AI).
Use Spot VMs when the workload is fault-tolerant and can be interrupted because they provide access to unused virtual machine capacity at a lower price, and use committed use discounts (CUDs) when usage is steady and predictable over one or three years because they give discounts for committing to minimum usage or spend. Sustained use discounts are a separate option that gives automatic Compute Engine discounts for consistent use without a commitment.
Last updated