Cloud Models and Use Cases

Public cloud

A cloud model answers a different question from a cloud service. The service says what you run; the model says where the resources live and who they belong to. There are three to tell apart — public, private, and hybrid — and the exam almost always tests them through a scenario rather than a definition.

The fastest way in is to ask one question of each scenario: who can buy these resources?

A public cloud is built, controlled, and maintained by a third-party cloud provider. Anyone who wants to purchase cloud services can access and use public-cloud resources. This general public availability distinguishes public cloud from private cloud.

Note what the definition does not mention: the size of the deployment, the industry, or whether the customer is large. A resource is public-cloud because a provider operates it and it is on sale to whoever wants it — not because many people happen to be using it.

The consequences follow from that ownership. Public cloud can scale without capital expenditure for additional infrastructure. Applications can be provisioned and deprovisioned quickly, and customers pay only for what they use.

There is a price for that. The tradeoff is that the customer does not have complete control over resources and security.

Public cloud is therefore a suitable choice when rapid provisioning, usage-based spending, or scaling without purchasing hardware is more important than complete infrastructure control.

Private cloud

A private cloud is a cloud environment used by a single entity. Its resources remain dedicated to one organization.

Location is not part of that definition, and this is the trap the exam sets most often. A private cloud may run in an on-site datacenter or in a dedicated off-site datacenter, including one operated by a third party. A third party can host it and it stays private, because the resources are still dedicated to one organization rather than generally available for purchase.

Private cloud gives the organization's IT team much greater control. It can also keep the organization's data from being collocated with other tenants' data.

These benefits come with greater cost, and the organization must purchase hardware for startup and maintenance.

Private cloud is therefore appropriate when dedicated resources, data separation, or complete control over resources and security outweigh the cost and maintenance responsibilities.

Hybrid cloud

A hybrid cloud uses public and private clouds in an interconnected environment. Both halves must be present, and they must be connected: two unrelated environments belonging to the same company are not a hybrid design.

What the interconnection buys is a decision. Hybrid lets an organization determine which applications and services run in each environment. That placement flexibility can help the organization control security, compliance, or legal requirements — a regulated dataset stays private while a public-facing front end runs on public resources.

The second use case is capacity rather than placement. A hybrid design can let a private cloud surge during increased temporary demand by deploying public-cloud resources.

The word that matters is temporary. The pattern keeps the private environment while adding public capacity for the demand spike; it does not migrate the workload. When the spike passes, the public capacity goes and the private environment is still there.

Hybrid cloud is therefore appropriate when a workload needs deliberate placement across both environments, or temporary public-cloud capacity alongside a private cloud.

Two failure modes are worth naming, because scenarios are usually built out of them. Moving the whole workload to public resources is not hybrid: the private half is gone. Absorbing a spike inside the private cloud alone is not hybrid either: the public half was never used. Hybrid requires both environments to take part.

The three models side by side

The models are easiest to hold apart as a table of tradeoffs rather than three paragraphs.

Read the "control" row carefully, because it is the row scenarios are built from. Complete control over resources and security is a private-cloud property. Public cloud states the opposite outright as its tradeoff, and hybrid does not grant control everywhere — it lets you choose which environment a workload sits in, and the control comes from that placement.

The same discipline applies to cost. "No capital expenditure to scale up" is a public-cloud property; "hardware must be purchased for startup and maintenance" is a private-cloud one. A scenario that offers both at once is describing hybrid, not a better version of either model.

Choosing a model

Start with the requirement that most directly separates the models. One requirement usually does all the work, and the rest of the scenario is decoration.

Applied to a scenario, that becomes three questions in order.

1. Is there a private environment at all? If not, hybrid is out, whatever else the scenario says.
2. Are the resources dedicated to one organization, or on sale to anyone? That separates private from public.
3. Do both environments have to take part — by placement or by surge? Only then is it hybrid.

Choose public cloud when broadly available provider-operated resources, quick provisioning, and avoiding capital expenditure drive the decision. Choose private cloud when one organization needs dedicated resources and accepts hardware cost and maintenance in exchange for greater control. Choose hybrid cloud when the organization must connect public and private environments, choose workload placement between them, or extend private capacity temporarily with public resources.

Key takeaways

• Public cloud uses third-party-operated resources that are generally available for purchase and supports rapid scaling without capital expenditure, at the cost of complete control over resources and security.
• Private cloud dedicates resources to one organization and provides greater control with greater cost and hardware responsibility, whether it runs on-site or in a dedicated off-site datacenter operated by a third party.
• Hybrid cloud interconnects public and private clouds so an organization can control placement or use public resources for temporary demand, keeping the private environment in place.
• Decide with the requirement that separates the models: general availability points to public, dedication to one entity points to private, and needing both environments points to hybrid.