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 MicroBank

These are proposed environments, not claims about an existing MicroBank deployment:

Do not classify the local OrbStack machine as private cloud just because its address is private. Do not choose hybrid merely because the diagram has more boxes. A public name such as microbank.learnwithsk.dev describes how a client addresses an endpoint; it does not establish the cloud model or reveal every backend.

Make and check a decision

Write a three-sentence MicroBank proposal: who uses it, how capacity is obtained, and who operates it. For this foundations checkpoint, the observed environment is your local Linux learning machine and its small fixtures. The full application placement is still a design.

For a future provider-operated environment, distinguish the proposed cloud model from service choices such as VM, managed runtime, or database. Those change the division of operating work, which the next bite examines.

Would a local source repository plus an unrelated public landing page establish hybrid cloud? No. Explain the distinct cloud environments and application/data integration if you claim hybrid. Would a public URL prove its database is publicly reachable? No. Audience, endpoint exposure, and deployment model are separate questions.

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.

Apply this to MicroBank

Classify one proposed MicroBank environment by who supplies capacity and how it is provisioned. Your manually managed OrbStack machine is a local lab. A later public-cloud proposal needs an explicit operating boundary. Describe a hybrid design only when there are distinct cloud environments and a real integration requirement; a local Git checkout beside a cloud diagram is insufficient.

Keep this observation in the MicroBank readiness project.