Projects & the Resource Hierarchy

Google ACE ACE · Setting up a cloud solution environment

By Abu Anns · Last updated 20 November 2026Exam version: ACE (exam guide live 30 June 2026)

Why the resource hierarchy matters

Every resource in Google Cloud — every virtual machine, storage bucket, database, and firewall rule — lives inside a project, and every project sits somewhere in a resource hierarchy. The hierarchy is the skeleton of a Google Cloud environment: it decides who can access what, which policies apply where, and how costs roll up. The exam tests it heavily because getting the hierarchy wrong at the start is expensive to fix later. A well-designed hierarchy makes access control and governance almost automatic; a flat one, with dozens of projects floating loose under an organization, turns every permission change into manual work.

The hierarchy has four levels, from top to bottom: organization, folder, project, and resource. Policies and permissions set at a higher level are inherited by everything below it. That inheritance is the single most important idea in this guide: an IAM role granted on a folder applies to every project inside that folder and every resource inside those projects, and an organization policy set at the organization level constrains every folder and project beneath it. When a question describes a requirement that applies to a whole department or a whole environment, the answer is almost always to act at the folder or organization level, not project by project.

Organization, folders, and projects

An organization is the root of the hierarchy and represents a company. It is created automatically when you set up Google Workspace or Cloud Identity with a verified domain, and a Google Cloud account can have at most one organization. The organization node is where company-wide policies live: organization policies that restrict what any project may do, and IAM roles for central administrators. If there is no organization — for example, a personal Gmail account creating projects directly — those projects have no organization node above them, and organization policies cannot be applied.

Folders sit between the organization and projects and are optional but strongly recommended. They group projects by department, team, environment, or any structure that matches how the company works. A common pattern is folders for environments (production, staging, development) crossed with folders for teams, so that a policy like "no public IP addresses in production" can be applied once to the production folder. Folders can be nested up to ten levels deep, though in practice two or three levels is plenty. A project can belong to only one folder (or directly to the organization), and moving a project between folders changes which inherited policies and permissions apply to it — the exam sometimes tests whether you know that moving a project can change its effective permissions.

A project is the fundamental unit of organization in Google Cloud. It is the boundary for several things at once: billing (a project is linked to one billing account), API enablement (APIs are enabled per project), quotas (most quotas are per project), and IAM (roles can be granted at project level). Resources in different projects are isolated from each other by default — a VM in one project cannot see a bucket in another unless access is explicitly granted. Best practice is one project per application or per environment of an application, so that access, billing, and quotas stay cleanly separated. Project IDs must be globally unique across all of Google Cloud and cannot be changed after creation, while project names can be changed and need not be unique. Every project also has a project number, assigned by Google, which never changes.

Organization policies

Organization policies (the Organization Policy Service) let administrators set guardrails on what resources are allowed to do. Unlike IAM, which controls who can act, organization policies control what is allowed at all — even an organization administrator cannot override them by granting an IAM role. Common examples include restricting which regions resources can be created in, requiring OS Login on VMs, disabling the creation of external IP addresses, and requiring uniform bucket-level access on Cloud Storage buckets. Policies are defined by constraints (such as "allowed values" lists or boolean on/off settings) and are attached to the organization, a folder, or a project; a policy set at a higher level is inherited downward, and a policy set at a lower level can refine it but cannot widen it. The exam often tests the distinction: if the requirement is "no one in the company may create resources outside europe-west2", that is an organization policy, not an IAM role.

Cloud Identity and user management

Cloud Identity is Google's identity service for organizations that do not use Google Workspace. It provides managed user accounts and groups under the company's domain, so that employees sign in with their work email and administrators can manage them centrally. Users can be created manually in the Admin console, or automatically through directory synchronization from an existing identity provider such as Microsoft Active Directory (using Google Cloud Directory Sync) or through federation. Groups are the key management tool: grant IAM roles to a group, and everyone in the group inherits them — when someone joins or leaves a team, you change the group membership, not dozens of IAM bindings. The exam expects you to prefer groups over individual user grants whenever a question involves a team.

Enabling APIs and assessing quotas

Before a project can use a Google Cloud service, that service's API must be enabled in the project — Compute Engine API for VMs, Cloud SQL Admin API for databases, and so on. Enabling an API is a per-project action done in the console or with the command gcloud services enable. A classic exam scenario is a deployment that fails with an error about an API not being enabled; the fix is to enable the API, not to change permissions. Some APIs also require billing to be enabled on the project before they can be used.

Quotas limit how much of a resource a project can use — the number of VM instances, CPUs, IP addresses, or API requests per minute. Quotas exist to protect both Google and the customer from runaway usage. When a deployment fails because a quota is exhausted, the answer is to check the quota in the console (IAM & Admin → Quotas) and request an increase, which Google reviews; increases are not automatic or instant. Quotas are per project and per region for most resources, so a quota problem in one region may not exist in another. The exam sometimes pairs quotas with billing: a quota increase request for a project without billing enabled will not help, because many quotas are higher — or only increasable — once billing is attached.

Setting up operations tooling

Part of setting up a cloud environment is provisioning the products in Google Cloud's operations suite — Cloud Monitoring, Cloud Logging, and related tools — so that the environment is observable from the start. In practice this means creating a Monitoring workspace scoped to your projects, deciding how logs will be routed and retained, and making sure the metrics and logs you will need during an incident are actually being collected. The exam guide lists this under environment setup, and the monitoring and logging guide later in this package covers the details. For this section, remember the principle: observability is set up when the environment is created, not bolted on after the first outage.

Advertisement

Exam traps to watch for

Practice questions

Test yourself on this guide. Answers are checked instantly, with a full explanation for each question.

Q1.A company has 40 projects organized under folders for each department. The security team wants every project in the Engineering folder to require OS Login on all Compute Engine VMs. What is the most efficient way to achieve this?

Report an error

Q2.A developer tries to create a Cloud SQL instance in a new project and receives an error saying the Cloud SQL Admin API is not enabled. What should the developer do?

Report an error

Q3.Which statement about the Google Cloud resource hierarchy is correct?

Report an error

Q4.A startup is setting up Google Cloud for the first time. They want employees to sign in with their company email addresses and be managed centrally, but they do not use Google Workspace. Which service should they use?

Report an error

Q5.An administrator moves a project from the Development folder to the Production folder. What is a likely consequence of this move?

Report an error

Q6.A company wants to guarantee that no one — including project owners — can create Compute Engine VMs with external IP addresses anywhere in the organization. Which mechanism enforces this?

Report an error

Q7.A team is deploying a large batch of VMs in us-central1, but the deployment fails partway through with a quota error, even though the same deployment succeeded last month in the same project. What is the most likely cause, and what should they do?

Report an error

Q8.A new employee joins the data engineering team. The team manages 12 projects, and every team member needs the same set of permissions in all of them. What is the best way to give the new employee access?

Report an error