StayTalentReady

Cloud Computing

Week of 2026-09-29 · Download .docx

Objectives

Key terms

Public cloud
Multi-tenant infrastructure owned by a provider (AWS, Azure, GCP) and shared among customers.
Private cloud
Dedicated infrastructure operated exclusively for one organization.
Hybrid cloud
Combination of private on-premises infrastructure with one or more public cloud providers.
IaaS
Infrastructure as a Service — virtual machines, storage, and networking; customer manages OS and above.
PaaS
Platform as a Service — managed runtime and middleware; customer manages only application code and data.
FaaS
Function as a Service (serverless) — run individual functions in response to events with no server management.
Elasticity
Cloud property allowing automatic scaling of resources up or down based on real-time demand.
Shared responsibility
Security model splitting duties between provider (hardware, hypervisor) and customer (data, identity).
Type 1 hypervisor
Bare-metal hypervisor running directly on hardware without a host OS (e.g., VMware ESXi).
Type 2 hypervisor
Hosted hypervisor running as an application on top of an existing OS (e.g., VirtualBox).
Container
Isolated execution environment sharing the host OS kernel; packages app with its dependencies.
Snapshot
Point-in-time capture of a VM's complete state (memory + disk) for rollback.
3-2-1 rule
Keep 3 copies of data, on 2 different media types, with 1 copy stored offsite.
RTO
Recovery Time Objective — maximum acceptable downtime after a disaster.
RPO
Recovery Point Objective — maximum acceptable age of data lost after a disaster.

The concept

Cloud computing delivers on-demand IT resources over the internet on a pay-per-use basis, eliminating the upfront capital expenditure of purchasing and maintaining physical hardware. Understanding cloud models requires knowing both the deployment model (who owns and controls the infrastructure) and the service model (how much the customer manages).

Deployment models: A public cloud is multi-tenant, owned and operated by a provider, and billed by consumption. A private cloud is dedicated to one organization, offering more control but requiring capital investment. A hybrid cloud integrates private and public infrastructure, enabling cloud bursting — overflowing workloads to the public cloud during peak demand. A community cloud serves organizations with shared regulatory or mission requirements (e.g., government agencies sharing a FedRAMP-authorized cloud).

Service models define the customer's management boundary. In IaaS, the provider manages hardware and hypervisor; the customer manages everything from the OS upward. In PaaS, the provider also manages the OS, runtime, and middleware; the customer only deploys application code and data. In SaaS, the provider manages everything; the customer manages only their data and user accounts. FaaS (serverless) takes this further — customers deploy individual functions and the provider manages all execution infrastructure. The shared responsibility model ensures both parties know exactly what each must secure.

Elasticity is cloud's defining operational advantage: resources scale automatically in response to actual demand, then scale back when demand drops — no over-provisioning needed.

Virtualization runs multiple isolated virtual machines (VMs) on a single physical server. Type 1 (bare-metal) hypervisors like VMware ESXi install directly on hardware for maximum efficiency. Type 2 (hosted) hypervisors like VirtualBox run as applications on a host OS, ideal for development laptops. Containers (Docker) share the host OS kernel and package an application with only its dependencies — faster to start and more resource-efficient than full VMs, but providing less isolation.

Cloud storage types serve different workloads. Object storage (S3, Azure Blob) holds unstructured data at virtually unlimited scale using unique object IDs. Block storage presents virtual disks to VMs. The 3-2-1 backup rule is the industry standard: three copies, two media types, one offsite — protecting against hardware failure, media degradation, and site-level disasters. RTO defines how quickly systems must be restored; RPO defines how much data loss is acceptable.

Worked examples

Example 1: A startup launches a web app on a single EC2 instance (IaaS). Black Friday traffic overwhelms it. The team attaches an Auto Scaling Group — when CPU exceeds 70%, it automatically launches additional identical instances; when traffic drops, it terminates them. This is cloud elasticity in action, paying for capacity only while it is needed.
Example 2: A sys admin takes a snapshot of a production VM before applying a major OS update. The update breaks the application. The admin reverts to the snapshot, restoring the VM to its exact pre-update state in minutes. Without the snapshot, restoring from backup would take hours. This is why snapshots are taken before every significant change.

Common mistakes

Self-check

Try each one before you look. A miss here costs nothing and tells you exactly what to reread.

1. Which cloud service model requires the customer to manage the operating system?
2. What does the 3-2-1 backup rule specify?
3. Which hypervisor type runs directly on server hardware without a host OS?
4. What is the primary advantage of containers over virtual machines?
5. RPO (Recovery Point Objective) defines:

Canvas is the official record. This companion enhances the PGCC curriculum; it does not replace it. Last name and class year only. Students with a 504 plan or IEP: your accommodations apply.

↑ Back to top