Prepare and test your skills

Prepare and test your skills

Discussing the benefits and tradeoffs of Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) means comparing control, operational work, scalability, and flexibility. The central tradeoff is control versus abstraction: more control gives more flexibility but requires more management, while more abstraction reduces work but limits customization.
Infrastructure as a Service (IaaS) is the cloud model that gives the customer the most control over virtualized computing resources. In IaaS, the customer manages the operating system, middleware, and applications, while Google manages virtualization, servers, storage, and networking. Google Cloud IaaS examples include Compute Engine (opens in a new tab) and Google Cloud VMware Engine.
Compute Engine creates virtual machine and bare metal instances with choices of machine series, storage, accelerators such as graphics processing units (GPUs) or tensor processing units (TPUs), and guest operating systems. Google Cloud VMware Engine runs VMware workloads natively on Google Cloud without re-architecting, which supports modernization. The tradeoff is high operational overhead because the customer must patch, configure, secure, and maintain everything above the virtualization layer, and cost efficiency depends on the resources selected within quotas.
Managed platforms that act as Platform as a Service (PaaS) balance control with ease of use by having Google run the platform. In PaaS, Google manages infrastructure, operating system, and runtime or control plane, so the customer focuses on code, containers, and jobs. Examples include Batch, Cloud Run, Cluster Director, Cluster Toolkit, and Google Kubernetes Engine (GKE).
Batch schedules, queues, and runs large batch jobs by provisioning Compute Engine resources and managing their lifecycle. Cloud Run runs containers that respond to web requests or events without managing servers or clusters, and Cluster Director automates setup of artificial intelligence, machine learning, or high performance computing clusters. GKE manages the Kubernetes control plane while the customer deploys container workloads, with Autopilot mode offering more abstraction and Standard mode offering more infrastructure control. The tradeoff is less ability to customize the operating system and runtime in exchange for built-in scalability, high availability, and resilience, such as regional clusters surviving zone outages.
Software as a Service (SaaS) is the cloud model where the provider runs the complete application for the customer. Terraform Cloud is a SaaS example that runs Terraform in a stable, remote environment and securely stores state and secrets. Data flows from customer code and version control into the hosted application, and the customer uses the result through the application or its application programming interface.
Terraform Cloud integrates with the Terraform command-line interface and connects to GitHub, GitLab, and Bitbucket, so new commits can automatically trigger plans. SaaS offers the lowest operational overhead because the provider manages the application environment. The tradeoff is limited control compared with IaaS or PaaS, since the customer depends on the provider for features and runtime, while a self-hosted option such as Terraform Enterprise returns runtime control to the customer.
The three models differ across who manages each layer and how much daily work remains with the customer. IaaS keeps the most work with the customer, PaaS splits platform work with Google, and SaaS shifts application operation to the provider. Infrastructure as Code (opens in a new tab) reduces manual work across all three models by defining desired state in versioned files that create reproducible development, test, and production environments.
| Model | Customer manages | Operational overhead | Scalability and flexibility |
|---|---|---|---|
| IaaS | Operating system, middleware, and applications | High | Configurable virtual machines and bare metal; choice of machine series, storage, GPUs, TPUs, and guest operating systems |
| PaaS | Code, containers, jobs, and workload configuration | Lower; Google manages lifecycle, control plane, or all infrastructure | Cloud Run is highly scalable; Batch runs jobs at scale; GKE regional clusters resist zone outages |
| SaaS | Use of the application and connection to code | Lowest; provider manages application environment | Provider-managed scale; flexibility limited to application features |
Exam tip: IaaS requires you to manage the operating system, middleware, and applications; managed platforms shift operating system and runtime management to Google, while SaaS shifts the entire application to the provider.
Defining IaaS, PaaS, and SaaS means explaining how each model divides management responsibility between the customer and the provider. Google Cloud describes the models as a spectrum where IaaS gives the most control and SaaS gives the least operational work. More control means more setup and upkeep, while more abstraction means faster deployment with less maintenance.
Infrastructure as a Service (IaaS) is the cloud service model that offers the most control over the computing environment. In IaaS, the customer manages the operating system, middleware, and applications, while Google manages virtualization, servers, storage, and networking. Google Cloud IaaS solutions include Compute Engine (opens in a new tab), which creates and runs virtual machine and bare metal instances, and Google Cloud VMware Engine, which runs VMware workloads natively while Google manages hardware and the VMware platform.
IaaS lets teams build custom environments without buying physical data centers. It best supports workloads needing deep customization and is often chosen by infrastructure architects, machine learning infrastructure engineers, and platform engineers. The customer also owns patching, scaling decisions, and application configuration above the provider foundation.
Managed platforms and services that form Platform as a Service (PaaS) let customers ship code while Google runs the platform. In PaaS, Google manages infrastructure, operating system, and runtime environments so customers focus on workloads rather than servers. The workload flows from customer code or containers to Google-managed infrastructure, which provisions resources, runs the job, and manages its lifecycle.
Google Cloud managed solutions include peer options with different levels of abstraction:
Control still varies inside PaaS. GKE Standard mode gives more infrastructure control than Cloud Run, which abstracts away all infrastructure, and GKE Autopilot mode abstracts more than Standard mode.
Software as a Service (SaaS) is the cloud service model where the provider delivers a complete, ready-to-use application over the internet. In SaaS, the customer does not manage infrastructure, operating system, runtime, or application code and simply uses the application. Data flows from existing customer tools into the hosted application, and the customer signs in and uses features through the application or its interface.
In Google Cloud documentation, Terraform Cloud is a SaaS application that runs Terraform remotely, stores state and secrets, and connects to version control so new commits can trigger plans. Cloud Identity (opens in a new tab) is an identity as a service platform that centrally manages users and groups and enables single sign-on with third-party SaaS applications. SaaS best supports tasks where a ready-made capability solves the problem without building servers or platforms.
Comparing the three models shows how responsibility, control, and typical users change across the spectrum. IaaS keeps operating systems and applications with the customer, PaaS keeps only code and data with the customer, and SaaS keeps only usage and settings with the customer.
| Model | What the customer manages | What the provider manages | Example services |
|---|---|---|---|
| IaaS | Operating system, middleware, applications | Virtualization, servers, storage, networking | Compute Engine, Google Cloud VMware Engine |
| PaaS | Workloads and application code | Infrastructure, operating system, runtime environments | Batch, Cloud Run, Cluster Director, GKE |
| SaaS | Nothing except use of the application | Everything, including the application itself | Terraform Cloud, 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.
Continue with this study guide: Define IaaS, PaaS, and SaaS.
Comparing benefits and tradeoffs across total cost of ownership (TCO), flexibility, shared responsibility, management level, and staffing means asking who does the work and what that work costs. IaaS offers the most customization but requires more people and operations, while managed platforms and SaaS lower long-term overhead by limiting low-level choices.
Total cost of ownership (TCO) is the full cost of running a workload, not just the hourly price of cloud resources. TCO includes provisioned resource costs plus personnel expenses and overhead to design, build, and maintain the deployment. The Google Cloud Pricing Calculator estimates resource costs, while fair comparison uses cost per unit of work, such as users served or requests processed.
Other TCO drivers increase cost beyond a single virtual machine:
Deployment scope links availability directly to cost. Zonal designs cost less and are simpler, while regional and multi-regional designs add storage, transfer, and database synchronization work. A global load balancer requires Premium Tier networking, while a regional load balancer can use Standard Tier networking.
Flexibility and cost tradeoff with Compute Engine (opens in a new tab) shows why IaaS gives the most choice at the highest management price. The customer controls the operating system, middleware, and applications while Google manages virtualization, servers, storage, and networking. That freedom supports choices of machine series, storage services, accelerators, and guest operating systems, but requires long-term budgeting for patching, securing, scaling, and specialized staff.
Managed platforms trade some customization for lower overhead and more predictable operations. Google manages infrastructure, operating system, and runtime so teams focus on workloads instead of servers. Batch provisions Compute Engine resources for large jobs without direct virtual machine management, Cloud Run runs containers on serverless infrastructure without managing servers or clusters, and GKE manages the Kubernetes control plane while customers deploy workloads. Autopilot abstracts more infrastructure than Standard, so Standard preserves more control at the cost of more management.
Staffing and technical skill requirements shrink as Google manages more layers of the stack. IaaS demands deep infrastructure skills because the customer owns the operating system and middleware, while managed platforms let smaller or workload-focused teams operate without maintaining servers. Infrastructure as Code (opens in a new tab) further reduces manual effort by using versioned configuration files as a single source of truth.
| Model | Who manages operating system and runtime | Example services | Operational effort |
|---|---|---|---|
| IaaS | Customer manages operating system, middleware, and applications | Compute Engine, Google Cloud VMware Engine | Higher; team designs and maintains environment |
| Managed platforms | Google manages infrastructure, operating system, and runtime | Batch, Cloud Run, Cluster Director, GKE | Lower; platform absorbs infrastructure work |
IaaS is typically staffed by infrastructure architects, machine learning infrastructure engineers, and platform engineers who build environments from scratch. Managed-platform teams focus on application, data, and delivery skills. Complexity still grows with scope: zonal operation is simplest, regional and multi-regional designs add monitoring and careful rollouts, and hybrid and multicloud topologies need the most staff because tools and security controls differ across platforms.
Exam tip: Compare options by cost per unit of business value, such as requests processed, not by resource price alone, because operational effort is part of TCO.
Determining which computing model applies to a business scenario means matching control needs, team skills, delivery speed, and technical fit to IaaS, PaaS, or SaaS. The shared responsibility boundary sits in a different place for each model, so moving from IaaS to PaaS to SaaS reduces operational work but also reduces direct control.
Migrating legacy workloads with IaaS fits applications that depend on a specific operating system version, custom middleware, or server-level settings. A lift-and-shift migration moves the application to a cloud virtual machine with minimal changes, which avoids rewriting code and keeps migration fast. The provider manages physical data centers, networking, storage hardware, and virtualization, while the customer manages the operating system, middleware, runtime, applications, and data.
The tradeoff is that operational work remains with the organization. The team still patches the operating system, manages scaling, and handles backups for that server. Choose this path when preserving existing behavior matters more than reducing server management.
Building cloud-native applications with PaaS fits new development where speed to market matters. Developers write and deploy code while the platform supplies the runtime, scaling, and patching underneath, so the team spends time on features instead of infrastructure. Traffic flows from users to the provider-managed platform to the new application code without the team provisioning servers.
Google Cloud PaaS examples include App Engine for highly scalable applications, Cloud Run for containerized applications, and Cloud Run functions for event-driven code with zero server management. The tradeoff is less control over the runtime environment, so applications with unusual platform requirements may not fit. Choose PaaS when automation and efficiency matter but the team still needs to configure applications.
Deploying productivity tools with SaaS fits standard software such as email, document editing, and collaboration tools. SaaS prioritizes ease of use with the lowest administrative overhead. The provider manages hardware, operating systems, application code, updates, and storage, while the customer only configures settings and supplies data.
The software already exists and the provider runs it, so users consume it directly with no servers or application code for the organization to manage. This is the fastest path to value and the lightest demand on internal staff. The tradeoff is the least control, because the organization accepts the provider features, update schedule, and configuration options.
Migration approaches differ in how much change they require and how much cloud optimization they deliver. The six common approaches are peer options that align with different service models:
A four-step process flow for choosing IaaS, PaaS, or SaaS: identify control needs, assess staff and skills, weigh delivery speed, and confirm technical fit, with control needs pointing toward IaaS and skills or speed pointing toward PaaS or SaaS.
A four-step sequence for choosing a model applies control, skills, speed, and fit in order:
Exam tip: Lift-and-shift of an existing server-based application fits IaaS, a newly built application that should not require server management fits PaaS, and ready-made software consumed directly by users fits SaaS.
The cloud shared responsibility model divides security and management work between Google Cloud and the customer. Google Cloud secures physical facilities and the foundational platform, while the customer always secures data, access, and configuration. As you move from on-premises to IaaS, PaaS, and SaaS, more layers shift from the customer to Google Cloud.
The provider role in Google Cloud security and operations is to build, secure, and operate the trusted foundation that all workloads run on. Google Cloud protects data centers, hardware, networking equipment, and global infrastructure, and customers never manage physical security in Google Cloud. Every virtual machine, container, database, and application depends on this same physical foundation.
Foundational security controls are the default protections that Google Cloud applies for every customer. Google Cloud applies default encryption to data, maintains infrastructure controls for how data is gathered and stored, and enforces the Identity and Access Management (IAM) (opens in a new tab) policies that customers define. Enforcement flows from Google Cloud to customer workloads, so only approved identities can administer workloads or access data. Customer assignment decides who gets which role, while Google Cloud enforcement makes that decision stick.
Provider responsibility grows across service models. In IaaS, Google Cloud focuses on underlying infrastructure and physical security. In PaaS, Google Cloud takes on the runtime plus additional platform controls, with application and IAM management shared. In SaaS, Google Cloud owns most security responsibilities for the application itself.
Differentiating responsibility across deployment models shows how control shifts step by step from the customer to the provider. On-premises infrastructure is the baseline where the customer controls every layer, including physical data center, hardware, networking, operating systems, applications, and data. Cloud models gradually transfer parts of that work to the provider.
Four side-by-side stacks for On-Prem, IaaS, PaaS and SaaS show responsibility shifting from customer to Google, with Compute Engine for IaaS and BigQuery for PaaS as examples.
| Model | Google Cloud manages | Customer manages |
|---|---|---|
| On-premises | Nothing | Everything, including physical security, hardware, networking, operating systems, applications, and data |
| Infrastructure as a Service (IaaS) | Physical security and underlying infrastructure | Operating systems, applications, data, IAM configuration, and virtual network settings |
| Platform as a Service (PaaS) | Infrastructure, operating systems, and runtime, plus shared application and IAM controls | Data security, client protection, and shared application and IAM configuration |
| Software as a Service (SaaS) | Infrastructure, platform, and application, including most security controls | Access controls and data stored in the application |
IaaS examples include Compute Engine, Cloud Storage, Cloud VPN, Cloud Load Balancing, and Cloud DNS, which fit lift-and-shift migration. PaaS examples include App Engine, Google Kubernetes Engine (GKE), and BigQuery, which fit building applications without managing servers. SaaS examples include Google Workspace and Marketplace applications, which fit when the enterprise needs to process workloads without building the application.
Customer responsibilities for configuration, governance, and data protection mean securing what the customer brings to the cloud while Google Cloud protects the cloud itself. The customer is always responsible for securing data, controlling who can access it, and correctly configuring applications and policies. Misconfigurations on the customer side can still cause unauthorized access or data loss.
Data protection is the customer duty that never shifts to the provider. The customer owns its data, decides where data may reside, classifies confidential data by sensitivity, and locks down access so only approved identities can read sensitive data. Identity configuration requires least privilege, so each identity receives only needed permissions, with dedicated service accounts for workloads and rotation of credentials and tokens.
Application and workload configuration means the customer secures source code, configuration files, container images, and running containers and pods. In GKE Standard, for example, the operating system underneath containers is managed by Google Cloud, but the operating system inside customer containers and the application code remain customer responsibilities. Governance adds risk assessment, organization policies, Assured Workloads (opens in a new tab) controls for regulated data, and continual monitoring for drift and threats.
Exam tip: In IaaS you manage the operating system and applications, in PaaS you share application and IAM controls with Google Cloud, and in SaaS you manage only access and your data.
Infrastructure as a Service (IaaS) gives the most control because the customer manages the operating system, middleware, and applications while Google manages virtualization, servers, storage, and networking. Platform as a Service (PaaS) shifts operating system and runtime management to Google so the customer focuses on code, containers, and jobs. Software as a Service (SaaS) shifts the entire application to the provider so the customer manages only settings, access, and data.
Choose Infrastructure as a Service (IaaS) for lift-and-shift migration of existing server-based applications that depend on a specific operating system version, custom middleware, or server-level settings. Choose Platform as a Service (PaaS) for newly built cloud-native applications where speed to market matters and developers should focus on features instead of servers. Choose Software as a Service (SaaS) for standard ready-made software such as email, document editing, and collaboration tools where users consume the application directly with the lowest administrative overhead.
Google Cloud secures physical facilities and the foundational platform, applies default encryption to data, and enforces the Identity and Access Management (IAM) policies that customers define. The customer always secures data, access decisions, and configuration, including classifying data, applying least privilege, and securing source code and workload configuration. As workloads move from on-premises to Infrastructure as a Service (IaaS) to Platform as a Service (PaaS) to Software as a Service (SaaS), more layers shift from the customer to Google Cloud.
Last updated