Cloud Computing Fundamentals
What to know
Cloud computing is on-demand delivery of compute, storage, and networking as metered services. Candidates must be fluent in the service models - IaaS (raw VMs, you patch the OS), PaaS (managed runtime, you deploy code), SaaS (finished application) - and the deployment models: public, private, hybrid, and community. Know the essential characteristics: self-service provisioning, broad network access, resource pooling/multi-tenancy, rapid elasticity, and measured service. Understand the shared responsibility boundary: the provider secures the physical layer and, for managed services, the platform; the customer secures data, identities, and configuration.
Traps in questions
- Elasticity versus scalability: elasticity is automatic, demand-driven scale in and out; scalability is the capacity to grow. A question describing a monthly manual resize is scalability, not elasticity.
- High availability versus fault tolerance versus disaster recovery: HA masks component failure with redundancy inside a region; DR restores service after a region-scale event and usually involves data loss (RPO) and downtime (RTO).
- CapEx versus OpEx: moving to cloud converts upfront capital spend to operating spend; questions often phrase this as "reduces upfront investment," not "eliminates cost."
- Multi-tenancy is a feature of public cloud pooling, not a security flaw by itself.
Worked example
A retailer expects tenfold traffic on a single sale day per year and wants to avoid buying servers for peak capacity. Rule of thumb: if demand is spiky and the workload is stateless, elasticity on public cloud wins; if the workload is steady and highly regulated, reserved capacity or private deployment may win. The answer here is elasticity - automatic scale-out for the event, scale-in after.
Rule of thumb
Match the model to the control you need: IaaS for OS-level control, PaaS to minimize operations, SaaS to minimize ownership.