Describe Cloud Concepts
Start with the pricing model
Most cloud-concept questions are really asking you to map a business requirement to a term. The foundation is the consumption-based model: you rent capacity and pay only for what you use, which converts capital expenditure (buying servers up front) into operational expenditure (a recurring bill). Economies of scale is why unit prices fall - the provider aggregates demand across millions of customers and passes some savings on.
Deployment models
Public cloud runs on provider-owned shared infrastructure; private cloud is dedicated to one organization (on-premises or hosted by a third party); hybrid connects on-premises assets to cloud services; multi-cloud uses two or more providers. Trap: "cloud plus a second cloud" is multi-cloud, not hybrid, and a privately hosted cloud is still private even if someone else's datacenter runs it.
Service types and the shared responsibility model
IaaS - you manage the OS and everything above it. PaaS - the provider manages OS, runtime and middleware; you manage apps and data. SaaS - the provider manages nearly everything. The customer's side of the line always includes data, accounts, devices and identity, even in SaaS. Questions love asking who patches the guest OS: in IaaS, you do.
Benefits worth memorizing as pairs
High availability (uptime), scalability (vertical = a bigger machine; horizontal = more machines), elasticity (automatic scaling to match demand), reliability (recovery from failure), predictability (performance and cost), plus security/governance posture and manageability.
Rule of thumb
If the scenario says "add more servers," answer horizontal scaling; if it says "give the server more RAM," answer vertical scaling; if it says "pay only when code executes," that signals serverless. If it says "no servers to manage but you still control code," that is PaaS, not IaaS.