Cloud Hosting Glossary

The vocabulary of cloud hosting, defined without the marketing gloss. Every term is written for the decision it actually affects: what you pay, how fast your site runs, and what breaks when traffic spikes. 76 terms and counting.

Hosting Models

Bare Metal Server

A single-tenant physical server with no virtualization layer, giving you direct access to the hardware.

Cloud Hosting

Hosting that runs your site or app across a pool of virtual servers instead of one fixed machine, so capacity can grow or shrink on demand.

Colocation

Renting rack space, power, and network in a data center for hardware you own and maintain yourself.

Dedicated Server

A physical machine rented entirely to you, with no other tenants sharing its CPU, memory, or disk.

Hyperscaler

One of the giant cloud providers (AWS, Google Cloud, Azure) operating global data centers and hundreds of services at massive scale.

IaaS

Infrastructure as a Service rents raw building blocks (compute, storage, networking) that you assemble and manage yourself.

Managed Hosting

Hosting where the provider handles server administration, security patching, updates, and backups so you don't have to.

PaaS

Platform as a Service runs your code from a git push without you touching servers, operating systems, or scaling config.

SaaS

Software as a Service is a finished application you access over the web and pay for by subscription, with all infrastructure hidden.

Serverless

A model where you deploy code and the provider runs it on demand, billing only for actual execution with no server to manage or keep running.

Shared Hosting

The cheapest hosting tier, where many websites share one server's resources with no guaranteed slice for any single site.

Unmanaged Hosting

Hosting where you get raw infrastructure and handle all administration yourself, from OS patching to security.

VPS

A Virtual Private Server is a dedicated slice of a physical machine with its own CPU, RAM, and root access, priced at a flat monthly rate.

Compute

Autoscaling

Automatically adding or removing server capacity in response to real-time demand, so you pay for what traffic actually requires.

Cold Start

The delay when a serverless function or idle container has to spin up from scratch before handling its first request.

Compute Instance

A single virtual server you rent from a cloud provider, defined by its vCPU count, memory, and disk.

Container

A lightweight, isolated package of an application and its dependencies that runs consistently across any host.

Kubernetes

An open-source system that automates deploying, scaling, and managing containerized applications across a cluster of servers.

Managed Kubernetes

A service where the provider runs and maintains the Kubernetes control plane, so you only manage your workloads and worker nodes.

RAM

Random-access memory holds the data your server is actively working with, and running out forces slow disk swapping or crashes.

Serverless Function

A single piece of code that runs on demand in response to an event, billed only for its execution time.

Shared vs Dedicated CPU

Shared CPU plans borrow processing power from a common pool and can slow under contention. Dedicated CPU guarantees full-time access to your cores.

vCPU

A virtual CPU is a share of a physical processor core assigned to your server, usually one hardware thread.

Storage

Block Storage

Raw disk volumes you attach to a server and format like a physical drive, used for databases, filesystems, and boot disks.

IOPS

Input/output operations per second: a measure of how many individual reads and writes a storage volume can handle, key for database performance.

Object Storage

Storage for unstructured files accessed over HTTP, priced per gigabyte and built to hold anything from backups to media at massive scale.

Persistent Volume

Storage that survives independently of the server or container using it, so data isn't lost when compute is destroyed.

Snapshot & Backup

Point-in-time copies of a server or volume you can restore from, the difference being that backups are stored separately for durability.

SSD, NVMe & HDD

The three storage media tiers: HDD (cheap, slow spinning disks), SSD (fast flash), and NVMe (fastest flash over a direct bus).

Databases

Cache

A fast, in-memory store that holds frequently accessed data so your app can skip slower database or API calls.

Connection Pooling

Reusing a small set of open database connections across many requests, instead of opening a new one each time, to avoid exhausting the database.

Managed Database

A database the provider runs for you, handling provisioning, patching, backups, replication, and failover so you only use it.

NoSQL Database

A database that trades rigid tables and schemas for flexible data models (documents, key-value, graph), often to scale writes horizontally.

Read Replica

A copy of your database that stays in sync with the primary and serves read queries, spreading load and adding redundancy.

SQL Database

A relational database that stores data in structured tables with enforced schemas and is queried with SQL, like PostgreSQL or MySQL.

Networking & Delivery

API Gateway

A front door for your APIs that handles routing, authentication, rate limiting, and traffic shaping in one place.

Bandwidth

The volume of data transferred to and from your server, often metered and billed, especially for outbound traffic.

CDN

A network of servers spread around the world that caches your content close to users, cutting load times and origin bandwidth.

DNS

The system that translates human-readable domain names into the IP addresses servers actually use to find each other.

Edge Computing

Running code at locations physically close to users rather than in one central data center, to cut latency for global audiences.

Egress Fees

Charges for data leaving a provider's network, a major and often surprising cost on hyperscaler clouds.

Ingress

Data flowing into a provider's network, which is typically free, in contrast to the egress it costs to send data out.

IPv4 vs IPv6

The two versions of internet addressing: IPv4 addresses are scarce and increasingly billed for, while IPv6 is abundant and usually free.

Latency

The time delay between a request and its response, driven heavily by physical distance between user and server.

Load Balancer

A service that distributes incoming traffic across multiple servers, improving capacity and keeping the site up if one server fails.

Message Queue

A buffer that passes tasks between parts of a system asynchronously, so slow work happens in the background instead of blocking a request.

VPC

A private, isolated network within a cloud provider where your resources communicate securely without exposing traffic to the public internet.

Pricing & Billing

Free Tier

A permanently or temporarily free allowance of a service, used to let you build and test before paying, with limits that cap real production use.

Hidden Costs

Charges beyond the advertised base price, like egress fees, IP addresses, backups, and support, that inflate the real bill.

Introductory Pricing

A discounted rate for your first billing term that jumps to a much higher renewal price later, common in shared and managed WordPress hosting.

Metered Billing

Billing that measures and charges for exact resource consumption, the mechanism behind pay-as-you-go pricing.

Overage Fees

Charges that kick in when you exceed a plan's included allowance for bandwidth, storage, requests, or compute.

Pay-as-you-go

Billing based on actual consumption (compute time, bandwidth, requests) rather than a fixed monthly fee, so costs scale with usage.

Renewal Pricing

The price you pay after an introductory term ends, which is the true long-term cost of a hosting plan.

Reserved Instance

A discount in exchange for committing to a server for one to three years upfront, cheaper than on-demand for steady workloads.

Spot Instance

Spare cloud capacity sold at a steep discount that the provider can reclaim on short notice, ideal for interruptible workloads.

Total Cost of Ownership

The full cost of running a workload, including infrastructure, add-ons, support, and the engineering time to operate it, well beyond the sticker price.

Reliability & Performance

Availability (the nines)

Shorthand for uptime levels by counting nines: 99.9% is three nines, 99.99% is four, each nine cutting allowed downtime roughly tenfold.

Availability Zone

An isolated data center within a region, with its own power and networking, so failure in one zone doesn't take down the others.

Data Center Tier

A rating from Tier 1 to Tier 4 describing a data center's redundancy and expected uptime, with higher tiers tolerating more failures without going down.

High Availability

Designing a system with no single point of failure, so it keeps running even when individual components fail.

Redundancy & Failover

Running duplicate components so that if one fails, another takes over automatically, the mechanism behind high availability.

Region

A geographic location where a provider operates data centers. Choosing one near your users cuts latency, and spreading across them adds resilience.

SLA

A formal promise of a service's availability, usually with financial credits owed to you if the provider misses it.

Uptime

The percentage of time a service is available and working, the headline reliability number hosts advertise.

Security & Compliance

Compliance Standards

Formal certifications and regulations (SOC 2, ISO 27001, PCI DSS, HIPAA, GDPR) that govern how providers handle security and data, and which you may be required to use.

DDoS Protection

Defenses that absorb or filter distributed denial-of-service attacks, where floods of traffic try to overwhelm your server.

Encryption at Rest & in Transit

Protecting data both while it's stored (at rest) and while it moves across networks (in transit), so it's unreadable if intercepted or stolen.

Firewall

A filter that controls which network traffic can reach your server, blocking everything except the ports and sources you allow.

Identity & Access Management

The system for controlling who can access what in your cloud account, granting each user and service only the permissions it needs.

Secrets Management

Securely storing and controlling access to credentials like API keys, database passwords, and tokens, instead of hardcoding them.

SSL/TLS Certificate

The certificate that enables HTTPS, encrypting traffic between browser and server and proving the site is who it claims to be.