Akamai Cloud Computing Foundations key terms

Delivery of computing resources such as servers, storage, and networking over the internet on demand, metered by usage, so teams avoid purchasing and maintaining physical hardware in their own facilities.

A virtual server with allocated CPU, memory, disk, and network interfaces that runs on shared physical hardware, created from an image and managed through a control panel, API, or CLI.

A pricing tier where virtual CPUs share physical cores with other customers' workloads, keeping cost low for light or bursty applications but allowing performance to vary when neighbors are busy.

A pricing tier that assigns physical processor threads exclusively to one instance, giving predictable performance for latency-sensitive or sustained workloads such as production databases and application servers.

An instance class provisioned with a large ratio of memory to CPU, suited to workloads that keep working sets in RAM, such as in-memory caches, analytics jobs, and busy databases.

Network-attached disk storage that attaches to an instance and can be resized, detached, and reattached independently of the instance's boot disk, making it suitable for application and database data.

A service for storing unstructured data as objects inside buckets, reachable over HTTP APIs rather than mounted as a filesystem, commonly used for media assets, logs, and backup archives.

A named container within object storage that holds objects and carries its own access settings; bucket names must be unique within the service namespace and are tied to a chosen location.

A stored template of an operating system and disk contents used to create new instances, sourced from the provider, the community, or a disk you captured yourself.

A point-in-time copy of a volume or disk kept separately from the original, used to restore data or to build new volumes and images after accidental deletion or corruption.

An automated, scheduled copy of an instance's disks retained by the platform on a rolling basis, letting you restore the whole machine to an earlier state after failure or mistakes.

A managed service that runs the Kubernetes control plane and worker nodes for you, so teams deploy containerized applications without building or patching cluster infrastructure themselves.

A group of worker nodes inside a Kubernetes cluster that share the same plan and size, letting you scale capacity by changing the pool's node count.

A managed load balancer that spreads incoming TCP or HTTP requests across several backend instances, checks their health, and can terminate TLS before traffic reaches your servers.

A private layer-2 network you create inside a region and attach instances to, letting them exchange traffic on isolated addresses that are not routable from the public internet.

An address used on a private network segment such as a VLAN, carrying instance-to-instance traffic that stays inside the provider's network and is unreachable from the internet.

A globally routable address assigned to an instance so internet clients can reach it; it can be reserved so the address survives rebuilding or replacing the instance.

A public IPv4 address held in your account independently of any single instance, so you can move it to a replacement machine without updating DNS records.

A hosted authoritative DNS service where you create zones and records for your domains, with the provider's name servers answering queries and full control available through the API.

A globally distributed caching layer that serves content from edge locations near users, cutting latency and origin load while absorbing traffic spikes and some attack traffic.

A packet filter attached to instances or node pools, with inbound and outbound rules that permit or deny traffic based on protocol, port, and source or destination address.

The collection of user accounts, roles, and permission grants that determines who may view or change which resources, supporting least-privilege access for people and automation.

A named bundle of permissions that can be granted to users or tokens, describing which resource types and actions they are allowed to use.

A scoped credential used to authenticate programmatic calls to the provider's API, with selectable permissions and an optional expiry so automation receives only the access it needs.

A matched public and private key used instead of a password to log in to Linux instances; the public key is placed on the server while the private key stays with the user.

A login safeguard that requires a second proof beyond the password, such as a time-based code from an authenticator app, reducing damage when a password is stolen.

Traffic scrubbing that detects and absorbs volumetric and protocol attacks before they reach your instances, keeping services reachable without manual action during an attack.

A digital certificate that encrypts traffic between clients and your service and proves the server's identity; certificates are uploaded to load balancers and renewed before they expire.

The browser-based control panel where you create and configure instances, volumes, networks, and billing settings, and review usage and monitoring data.

A documented HTTP interface through which software creates, reads, updates, and deletes cloud resources, forming the base layer for tooling, scripts, and infrastructure-as-code.

A command-line program that wraps the provider's API, letting you script routine tasks such as launching instances or listing volumes from a terminal or CI job.

A reusable script stored by the platform that runs automatically the first time an instance boots, installing packages and configuring software consistently across deployments.

A boot-time configuration mechanism that reads user data supplied at instance creation, letting you define users, packages, and files declaratively without building a custom image.

A catalog of preconfigured one-click applications and stacks that deploy onto new instances, shortening setup for common software such as web servers, databases, and monitoring tools.

A geographic location containing one or more data centers where resources can be placed; the choice affects latency to users and which plans and services are offered.

A label attached to resources that groups and filters them by purpose, environment, or owner, simplifying inventory checks and cost review across a growing account.

A pricing model where charges accrue for each hour a resource exists, capped at the plan's monthly price, so short-lived resources cost proportionally less.

The monthly quantity of outbound network traffic included with your services at no extra charge; usage above the allowance is billed per gigabyte.

A paid subscription that sets response-time targets and contact channels for help requests, ranging from standard ticket support to round-the-clock access for critical incidents.