Section 36: Cloud computing

What is Cloud Computing?

A model for enabling ubiquitous, convenient on-demand network access to a shared pool of configurable computing resources.

What is the Hybrid Cloud deployment model?

Two or more distinct cloud infrastructures that remain unique entities, bound together by standardized technology.

What is the Private Cloud deployment model?

Cloud infrastructure provisioned for exclusive use by a single organization comprising multiple consumers.

What is the Community Cloud deployment model?

Cloud infrastructure provisioned for exclusive use by a specific group of consumers from organizations with shared concerns.

What is the Public Cloud deployment model?

Cloud infrastructure provisioned for open use by the general public.

What is a Bare Metal server?

A server where the OS runs directly on the hardware, with no hypervisor.

What level of access does the customer get with IaaS?

Operating system level.

With PaaS, who manages the operating system, and who manages the applications/data?

The provider manages the OS; the customer manages applications and data.

With SaaS, who manages everything from the facility up to the data?

The provider.

What does Clustering support?

Combining multiple physical systems into a single virtual system.

How does Private Cloud differ from Public Cloud in terms of who it serves?

It works the same way, but services are provided to internal business units instead of external public enterprises.

What does Virtualization support?

Running multiple virtual systems on a single physical system.

What are the five NIST characteristics that define cloud computing?

On-demand self-service, rapid elasticity, broad network access, resource pooling, and measured service.

What does "rapid elasticity" mean in cloud computing?

Resources can be quickly scaled outward (up) or inward (down) to match demand, often automatically — the consumer only pays for what's currently provisioned.

What does "resource pooling" mean in cloud computing?

Multiple customers share the same underlying physical hardware (multi-tenant), with resources dynamically assigned/reassigned — this lowers costs compared to dedicated hardware per customer.

What does "measured service" mean in cloud computing?

Usage is metered (CPU, storage, bandwidth, etc.) so billing reflects actual consumption — typically an ongoing operational expense rather than upfront capital expense.

Why is running multiple applications directly on the same OS (without virtualization) considered bad practice?

A problem with one application can crash the others, since they all share the same OS instance — virtualization isolates each into its own VM so a failure doesn't cascade.

What's the key architectural difference between a container and a virtual machine?

A VM includes its own full OS. A container does NOT include an OS — only the application's dependencies (executables/libraries) and the app itself, making containers much smaller and faster to provision.

What is Docker, and where does it sit in the virtualization stack?

The most well-known container engine — sits on top of a host OS (typically Linux), with containers running on top of it.

What is VRF, and what does it allow on a single physical router?

Virtual Routing and Forwarding — allows separate, isolated routing tables for different customers/departments on the same physical router, commonly used for MPLS Layer 3 VPNs.

Can customers be given direct administrative access to their own separate VRF, the way they can with virtualized firewall contexts?

No — VRF uses a single shared configuration on the router, so individual customers can't be given isolated admin access the way separate firewall security contexts allow.

In the data center stack (facility → network → storage/compute → hypervisor → OS → applications → data), where is the responsibility boundary for IaaS?

The provider manages everything through the hypervisor. The customer manages everything from the OS upward (OS, applications, data).

Why isn't a "dedicated server" from a Public Cloud provider actually Private Cloud, even though the hardware is dedicated to one customer?

Because the surrounding infrastructure (switches, routers, storage systems) is still shared with other customers — true Private Cloud requires the infrastructure to be exclusively for one organization.

What is "cloud bursting," and what's an example use case?

Scaling out from a Private Cloud into Public Cloud when internal capacity is exceeded — e.g. using Public Cloud as a disaster recovery site for an organization's own Private Cloud data center.

Does adopting cloud computing eliminate the need for in-house IT/network engineering staff?

No — customers (especially with IaaS) still need in-house staff to design the network, and manage everything from the OS level upward; the provider only manages the underlying infrastructure.