CLF key terms

On-demand delivery of computing resources such as servers, storage, and databases over the internet, with pay-as-you-go pricing instead of buying and maintaining physical hardware yourself.

The ability to automatically acquire more resources during demand spikes and release them when demand drops, so capacity tracks actual workload needs rather than peak guesses.

The capacity of a system to grow to handle increased load by adding resources, either vertically by enlarging one machine or horizontally by adding more machines.

Adding more individual instances to a pool so work is shared across them, often combined with a load balancer; also called scaling out.

Increasing the size or power of a single resource, such as moving to a larger instance type with more CPU and memory; also called scaling up.

Designing a system so it keeps serving requests despite component failure, typically by running redundant resources across multiple locations and health-checking them continuously.

The ability of a system to continue operating correctly when one or more components fail, usually through built-in redundancy with no interruption in service.

CapEx is upfront spending on long-lived assets such as data center hardware; OpEx is ongoing usage-based spending. Cloud shifts most IT costs from the first model to the second.

Cost advantages gained by purchasing and operating at very large volume, letting a cloud provider charge lower per-unit prices than a typical organization could achieve alone.

A full comparison of all costs of an option over time, including hardware, power, facilities, staffing, and maintenance, not just the visible purchase price.

A separate geographic area containing multiple isolated data center groups, chosen to place workloads near users and to satisfy data residency or latency requirements.

One or more discrete data centers with independent power, cooling, and networking inside a region, connected to sibling zones by low-latency links for redundancy.

A site closer to end users than a region, used to cache content and terminate connections so latency is reduced; many more of these exist than regions.

The division of security duties between the cloud provider, which secures the cloud itself, and the customer, who secures what they put in the cloud.

The service and practice of creating users, groups, and roles and granting them precisely scoped permissions through policies that allow or deny specific actions on resources.

Granting only the permissions a person or service needs to do its job, and no more, to limit the blast radius if credentials are misused.

An identity with permission policies that trusted entities can temporarily assume, providing short-lived credentials instead of long-term access keys.

Requiring two or more independent proofs of identity at sign-in, such as a password plus a one-time code, so a stolen password alone is not enough.

Protecting stored data by encoding it so it is unreadable without the correct key, covering volumes, object storage, and databases even if physical media is compromised.

Protecting data while it moves between systems by using protocols such as TLS, so intercepted traffic cannot be read or silently altered.

A stateful virtual firewall attached to resources that permits only explicitly allowed inbound and outbound traffic, evaluated by rules you define.

A stateless subnet-level filter with numbered allow and deny rules evaluated in order, providing a coarse extra layer beyond per-resource firewalls.

Meeting external laws, regulations, and industry standards, demonstrated through audits, reports, and controls that a provider and customer share responsibility for.

A logically isolated network you define, with your own IP address range, subnets, route tables, and gateways, hosting resources privately.

A range of IP addresses within a virtual network, placed in one availability zone and marked public or private to control routing to the internet.

A named hardware profile for a virtual server that fixes the CPU, memory, storage, and networking ratio, grouped into families optimized for different workloads.

Automatically adjusting the number of running instances to match demand using rules or schedules, improving availability while avoiding paying for idle capacity.

Distributing incoming traffic across multiple healthy targets and routing around failed ones, so no single server becomes a bottleneck or single point of failure.

A durability-and-cost tier for object storage, ranging from frequently accessed to archival, chosen per object based on how quickly retrieval is needed.

Rules that automatically transition objects to cheaper storage classes or delete them after defined periods, controlling cost without manual cleanup.

The probability that stored data is not lost or corrupted over time, expressed as expected annual failure rates and distinct from how quickly data can be retrieved.

A content delivery network that caches content at edge locations worldwide, reducing latency for users and offloading traffic from origin servers.

A model where you deploy code or managed services without provisioning or patching servers; the provider scales automatically and you pay for actual use.

A set of design pillars covering operational excellence, security, reliability, performance efficiency, cost optimization, and sustainability, used to review workloads against best practice.

Charging only for resources actually consumed, measured by time, requests, or data volume, with no upfront commitment or termination penalty.

A commitment to a specific instance family in a region for a one- or three-year term in exchange for a significant discount over on-demand rates.

A flexible discount commitment measured in dollars per hour of compute usage over a term, applying automatically to eligible usage across families and regions.

A tool for estimating the monthly cost of an architecture by configuring services, quantities, and regions before anything is deployed.

A console tool that visualizes historical spending and usage, filters by service or tag, and forecasts future costs from past trends.

Configurable thresholds that alert you, or trigger actions, when actual or forecast spending or usage exceeds the amounts you set.

A tiered subscription defining response times, channels, and advisory services for help requests, from basic self-service up to business-critical response targets.

An advisory service that inspects your account against best-practice checks for cost, security, fault tolerance, performance, and service limits, flagging issues.

A self-service portal for downloading compliance reports and agreements from the provider, useful as audit evidence for your own certification efforts.