Public, private, and hybrid cloud

Connect the machine to an operating model

You have compared virtualization, VMs, containers, and the Linux machine lifecycle. Now ask who supplies infrastructure and how consumers obtain it. No cloud account or provisioning is needed for this module.

Cloud computing includes on-demand access, network access, pooled resources, elasticity, and measured service. Virtualization can support this model, but one manually operated VM does not by itself establish a cloud.

These are simplified deployment distinctions grounded in NIST's cloud definition↗.

Unpack the cloud characteristics

On-demand self-service means a consumer can obtain resources through the service without a person manually handling every request. Resource pooling means the provider assigns shared capacity to consumers. Elasticity is the ability to expand or release capacity as demand changes. Measured service means usage is accounted for, supporting monitoring and often billing. Network access connects consumers to these capabilities.

Virtualization is one possible underlying mechanism; it is not the whole operating model. An administrator manually creating a single VM in an office may be doing useful virtualization without providing the self-service, pooling, and measurement expected of a cloud.

Work a classification example: an organization offers its own teams a self-service compute portal, quotas, and metered pooled capacity in a data center operated for that organization. That fits a private-cloud proposal even if an external company operates the hardware. If an unrelated provider offers infrastructure to the public, that is a public-cloud model even when the chosen workload has only private network access.

Avoid name-based shortcuts

A public-cloud service can have private network access. A private cloud can host a public website. “Public” describes the cloud deployment model, not whether every workload or storage object is open to everyone.

An on-premises rack is not automatically a private cloud; explain its provisioning and operating capabilities. Likewise, your OrbStack learning machine is a local environment, not evidence that you have built a private cloud.

Using more than one provider is generally described as multi-cloud. Hybrid emphasizes integration across distinct environments; the two labels can overlap. An unrelated laptop plus a public website does not demonstrate a working hybrid application.

Apply the distinctions to your learning site

These are proposed designs, not descriptions of LearnWithSK's current infrastructure:

Do not make hybrid the default just because it has more components. Your initial site is local and has no requirement for a split deployment. Public hosting remains optional later.

Make and check a decision

Write a three-sentence proposal for your learning site: who uses it, how capacity is obtained, and who operates it. For the current course setup, a sound answer is a local Linux learning machine created and managed by you for local practice. No claim of private cloud is needed.

For an optional future public static host, identify a provider-operated service offered to customers and a publishing workflow. That supports a public-cloud classification, but says nothing yet about who patches a runtime or whether PHP is supported. Those are service-model questions in the next bite.

Would a local source repository plus an unrelated public website alone establish hybrid cloud? No. Explain the integrated cloud environments and the application/data movement if you claim hybrid. Would a public website force its underlying infrastructure to be public cloud? No. Audience and deployment model are different dimensions.

Checkpoint

Classify a proposed environment using its ownership, access, and provisioning behavior. Explain why “runs in a VM,” “uses a private IP,” and “runs in our office” do not each determine a cloud model. Continue to service models to decide which responsibilities remain yours.