CDN Glossary: 50 Terms Every Developer Should Know

Content Delivery Networks involve specialized terminology spanning networking, caching, security, and performance engineering. This glossary defines the 50 most essential CDN terms that developers, DevOps engineers, and technical decision-makers encounter when evaluating, configuring, and optimizing CDN infrastructure. Each definition is written for practitioners who need precise, actionable understanding rather than marketing abstractions. Terms are organized by functional category for quick reference. For a comprehensive overview, see our guide on What Is a CDN? The Complete Beginner's Guide — FlyFish NWA.

What Are the Core Networking Terms in CDN?

Anycast

A network addressing method where the same IP address is announced from multiple geographic locations via BGP. Internet routing directs each user to the topologically nearest instance. CDNs use anycast for automatic, latency-based traffic routing without DNS manipulation. See our anycast vs unicast comparison.

Unicast

The standard one-to-one network addressing model where each IP address maps to a single server. Traditional hosting uses unicast; CDNs largely replaced it with anycast for edge delivery.

BGP (Border Gateway Protocol)

The routing protocol that governs how traffic flows between autonomous systems on the internet. CDNs announce their IP prefixes via BGP from each PoP to attract traffic to the nearest edge server.

DNS (Domain Name System)

The hierarchical system that translates human-readable domain names into IP addresses. CDNs use DNS resolution as one mechanism to direct users to the optimal edge server based on geography or server health.

Round-Trip Time (RTT)

The time in milliseconds for a network packet to travel from source to destination and back. CDNs reduce RTT by placing edge servers close to users, cutting round-trip distances from thousands of kilometers to tens.

Latency

The delay between a user action (such as clicking a link) and the start of the server response. Network latency is the physical transmission delay; CDNs minimize it through geographic proximity and protocol optimization.

Bandwidth

The maximum data transfer rate of a network connection, measured in bits per second. CDN edge servers aggregate massive bandwidth capacity across their networks, often exceeding 200 Tbps globally.

Throughput

The actual data transfer rate achieved, as opposed to theoretical bandwidth capacity. CDNs improve throughput through connection multiplexing, compression, and protocol optimization like HTTP/2 and HTTP/3.

Peering

A direct network interconnection between two autonomous systems for exchanging traffic without paying a transit provider. CDNs establish extensive peering relationships to reduce hops and costs.

IXP (Internet Exchange Point)

A physical facility where multiple networks connect to exchange traffic. CDNs deploy servers at IXPs to peer directly with ISPs, reducing latency and transit costs for end users.

What Caching Terms Should Every CDN User Understand?

Cache Hit

A request fulfilled directly from edge cache without contacting the origin server. Cache hits deliver the fastest response times and zero origin load.

Cache Miss

A request for content not present in edge cache, requiring a fetch from the origin server. The fetched content is then cached for subsequent requests.

Cache Hit Ratio (CHR)

The percentage of total requests served from cache. A CHR of 95% means 95 out of 100 requests hit cache. Well-optimized CDNs target 85-98% for static content. Explore caching strategies to improve this metric.

TTL (Time to Live)

The duration in seconds that cached content remains valid. Set via Cache-Control headers from the origin. Short TTLs (60-300s) suit frequently updated content; long TTLs (86400s+) suit versioned static assets.

Cache-Control

The HTTP response header that defines caching directives. Common values include max-age (TTL duration), no-cache (must revalidate), no-store (never cache), and s-maxage (CDN-specific TTL overriding max-age).

Cache Invalidation (Purge)

The process of removing cached content before its TTL expires. Triggered manually via API calls or automatically through content management systems when content updates.

Stale-While-Revalidate

A Cache-Control directive that allows the CDN to serve expired (stale) content while simultaneously fetching a fresh copy from the origin in the background. Improves perceived performance for frequently accessed resources.

Cache Key

The unique identifier the CDN uses to store and retrieve cached objects. Typically composed of the URL, query parameters, and optionally request headers. Misconfigured cache keys cause cache pollution or unnecessary misses.

Origin Pull

The CDN model where edge servers fetch content from the origin on-demand upon cache miss. The dominant model for most CDN configurations. Contrasts with origin push, where content is proactively distributed.

Cache Warming (Pre-fetching)

Proactively populating edge caches with content before users request it. Used before traffic events or after cache purges to prevent a flood of origin requests.

Surrogate Key

A tag attached to cached objects that enables selective purging of related content. For example, purging all objects tagged with product ID 1234 across all edge servers simultaneously.

Vary Header

An HTTP response header that tells CDNs to cache separate versions of a resource based on specific request headers such as Accept-Encoding or Accept-Language.

What Security Terms Are Essential for CDN Configuration?

DDoS (Distributed Denial of Service)

An attack that floods a target with traffic from many distributed sources to overwhelm its capacity. CDNs absorb DDoS traffic across their global network, preventing origin overload.

WAF (Web Application Firewall)

A security layer that inspects HTTP requests and blocks malicious traffic based on rule sets. CDN-integrated WAFs filter attacks at the edge before they reach the origin.

TLS/SSL (Transport Layer Security / Secure Sockets Layer)

Encryption protocols securing data in transit between client and server. CDNs terminate TLS at the edge, offloading cryptographic processing from the origin and reducing handshake latency.

TLS Termination

The process of decrypting TLS-encrypted traffic at the CDN edge rather than forwarding encrypted packets to the origin. Reduces origin CPU load and allows the CDN to inspect and cache content.

Origin Shield

A designated intermediate cache layer between edge servers and the origin. When multiple edge servers miss cache simultaneously, the origin shield consolidates requests to prevent origin flooding.

Hotlink Protection

A security feature that prevents other websites from embedding your CDN-hosted assets, such as images and videos, on their pages, consuming your bandwidth without authorization.

Rate Limiting

Restricting the number of requests a single client can make within a time window. Applied at the CDN edge to mitigate brute-force attacks, API abuse, and scraping.

Bot Management

CDN features that identify and classify automated traffic as legitimate bots (search engine crawlers) or malicious bots (scrapers, credential stuffers), applying different handling rules to each.

Token Authentication

A method where CDN-delivered content requires a cryptographic token in the URL, preventing unauthorized access to protected resources such as paid media files.

DNSSEC

DNS Security Extensions that add cryptographic signatures to DNS records, preventing DNS spoofing attacks that could redirect CDN traffic to malicious servers.

What Performance Terms Matter for CDN Optimization?

TTFB (Time to First Byte)

The time elapsed from the client sending an HTTP request to receiving the first byte of the response. CDNs reduce TTFB from hundreds of milliseconds to under 40ms for cached content by serving from nearby edge servers.

LCP (Largest Contentful Paint)

A Core Web Vital measuring when the largest visible element finishes rendering. CDN delivery of images and fonts directly improves LCP by reducing asset transfer time.

HTTP/2

A major revision of the HTTP protocol introducing multiplexed streams, header compression, and server push. Allows multiple requests over a single TCP connection, reducing page load times. See HTTP/2 and HTTP/3 in CDN.

HTTP/3 (QUIC)

The latest HTTP version built on UDP-based QUIC transport. Eliminates TCP head-of-line blocking, provides faster connection setup (0-RTT), and handles network migration for mobile users.

Brotli

A compression algorithm developed by Google that achieves 15-25% better compression ratios than gzip for text-based assets. Most CDNs support Brotli compression at the edge.

Image Optimization

Automatic CDN processing that converts images to modern formats (WebP, AVIF), resizes to match client viewport, and adjusts quality based on connection speed.

Connection Coalescing

An HTTP/2 feature where a single TCP connection handles requests to multiple hostnames sharing the same IP and TLS certificate. Reduces connection overhead for multi-domain CDN configurations.

Early Hints (103)

An HTTP status code that allows the CDN to send preload directives to the browser before the full response is ready, enabling parallel resource fetching during origin processing.

Prefetch

A browser hint instructing it to fetch resources that will likely be needed for the next navigation. CDNs can inject prefetch headers to speed up subsequent page loads.

TCP Optimization

CDN techniques including increased initial congestion window, TCP Fast Open, and optimized buffer sizing that improve throughput over long-distance and high-latency connections.

What Architecture Terms Define CDN Infrastructure?

PoP (Point of Presence)

A physical location housing CDN edge servers, networking equipment, and storage. Major CDNs operate 200-300+ PoPs worldwide. See our CDN architecture overview for how PoPs interconnect.

Edge Server

A server within a PoP that caches and delivers content to nearby users. Handles TLS termination, compression, and request routing. Learn more in our edge server deep dive.

Origin Server

The primary server hosting the authoritative version of website content. The CDN fetches content from the origin on cache miss and returns to it for dynamic requests.

Mid-Tier Cache

An intermediate cache layer between edge PoPs and the origin in hierarchical CDN topologies. Consolidates origin fetches from multiple edge servers in a region, reducing origin load.

Multi-CDN

An architecture using multiple CDN providers simultaneously for redundancy, performance optimization, or cost management. Traffic is distributed via DNS-based or real-user-measurement-based switching.

Edge Computing

Running application logic directly on CDN edge servers rather than the origin. Enables personalization, A/B testing, authentication, and data transformation at the network edge with minimal latency.

Reverse Proxy

A server that sits in front of the origin and intercepts client requests. CDNs operate as reverse proxies, caching responses and serving them to clients without exposing the origin directly.

Load Balancer

A component that distributes incoming traffic across multiple servers. CDNs use load balancing at multiple layers: globally across PoPs, and locally within each PoP across edge servers.

Health Check

Periodic probes sent from CDN edge servers to the origin to verify availability. Failed health checks trigger automatic failover to backup origins or stale cache serving.

Failover

The automatic process of switching traffic to a backup system when the primary system becomes unavailable. CDN failover operates at PoP level (rerouting to another PoP) and origin level (switching to backup origin).

Quick Reference: Most Commonly Confused Term Pairs

This glossary covers the foundational vocabulary for working with CDN infrastructure. For practical application, explore how CDNs work to see these concepts in action, or dive into the CDN architecture overview for a systems-level perspective on how these components fit together.