Akamai Web Performance key terms

A server in a content delivery network located close to end users that stores cached copies of content and serves requests on behalf of the origin, reducing latency and origin load.

The authoritative source server that holds the original, uncached version of content. Edge servers fetch from it on cache misses, so protecting and offloading it is a core CDN goal.

A physical data center location containing groups of edge servers and network equipment. Multiple PoPs spread geographically shorten the network path between users and content.

The time delay between a request being sent and a response arriving, measured in milliseconds. It is affected by physical distance, network hops, server processing, and protocol round trips.

The time required for a signal to travel from a client to a server and back. Each protocol handshake adds RTTs, so reducing round trips is a major performance lever.

The elapsed time from initiating a request until the first byte of the response reaches the client. It combines DNS, connection, TLS, and server processing delays.

The hierarchical system that resolves human-readable hostnames into IP addresses. DNS resolution is an early, often overlooked contributor to page load time and can be optimized with fewer lookups and faster authoritative responses.

A DNS record type that aliases one hostname to another. CDN onboarding commonly uses a CNAME so a customer's vanity hostname resolves to the provider's edge hostname.

A routing method where one IP address is announced from many locations, and network routing delivers each user to the topologically nearest one. It reduces latency and improves resilience.

A request served directly from an edge server's cached copy without contacting the origin. Hits lower latency and origin load and are the primary measure of caching effectiveness.

A request that cannot be satisfied from the edge cache, forcing a fetch from the origin or a parent tier. Misses increase latency and origin traffic and are reduced by proper cache keys and TTLs.

The percentage of requests (or bytes) served from cache rather than the origin. A higher ratio generally means better performance and lower origin cost, though byte-weighted and request-weighted ratios can differ.

The duration a cached object remains fresh before it must be revalidated or refetched. Longer TTLs raise hit ratios but risk serving stale content; shorter TTLs trade freshness for origin traffic.

An HTTP response header carrying directives such as max-age, s-maxage, no-cache, no-store, and private that govern how browsers and shared caches store and reuse content.

The storage layer on CDN servers that holds copies of content close to users. Its behavior is controlled by cache keys, TTLs, and header directives configured in the property.

The set of request attributes (such as hostname, path, query string, and selected headers or cookies) that uniquely identify a cached object. Overly broad keys cause wrong content; overly narrow keys hurt hit ratio.

The process of asking the origin whether a cached object is still valid, typically using conditional requests with ETag or Last-Modified. A 304 response refreshes freshness without resending the body.

An HTTP response header containing a validator token for a specific version of a resource. Clients send it back in If-None-Match to enable efficient conditional revalidation.

The proportion of traffic absorbed by the CDN rather than passed to the origin. It is closely tied to cache hit ratio and is a key indicator of CDN efficiency and origin cost savings.

Proactively requesting and caching objects before users ask for them, for example warming popular assets ahead of a launch. It improves first-user experience and smooths traffic spikes.

An operation that removes or marks stale cached objects at the edge so subsequent requests fetch fresh content. Purges can target URLs, cache tags, or whole properties and should be used sparingly.

A label attached to cached objects that lets operators purge many related objects at once without enumerating URLs. Tags simplify invalidation for content sets like a product catalog.

The core Akamai configuration object that defines how a hostname's traffic is handled, including routing, caching, and behavior rules. Properties are versioned and deployed to staging or production networks.

A hostname associated with a property that determines which traffic the configuration applies to. Hostnames must be validated and mapped before the property can serve production traffic.

An immutable snapshot of a property's configuration. Versions are created, validated, activated on staging or production, and can be rolled back, supporting safe change management.

A configurable rule within a property that changes how the edge handles matching requests, such as caching, compression, redirects, or header modification. Behaviors are applied through rules and match criteria.

A conditional block in a property configuration containing match criteria and a set of behaviors. Rules are evaluated in order, and the first matching rule typically determines the applied behaviors.

The conditions in a rule that determine whether it applies to a request, such as path patterns, hostnames, headers, cookies, or client characteristics. Precise criteria keep behavior application predictable.

The process of pushing a property version to the Akamai staging or production network. Activations propagate across many edge servers and take time, so changes should be staged and verified first.

A separate Akamai environment for testing a property version before production activation, reached via a special hostname. It lets teams validate behavior changes without affecting live users.

Reverting a property to a previously activated version when a change causes problems. Fast rollback capability is a key operational safeguard for configuration changes.

The provider-assigned hostname that customer traffic ultimately resolves to after DNS mapping. It identifies the edge network and may encode delivery options such as IPv6 or protocol support.

A binary, multiplexed version of HTTP that allows many concurrent streams over one connection and supports header compression. It reduces head-of-line blocking at the application layer compared with HTTP/1.1.

The version of HTTP that runs over QUIC and UDP rather than TCP. It reduces connection setup and handshake overhead and avoids TCP-level head-of-line blocking, improving performance on lossy networks.

A UDP-based transport protocol with built-in encryption and stream multiplexing. It combines transport and cryptographic handshakes to reduce connection establishment round trips.

The negotiation that establishes an encrypted session between client and server, including certificate verification and key agreement. It adds round trips before content can flow and can be optimized with session resumption and modern TLS versions.

A technique where the server attaches a signed, time-stamped certificate revocation status to the TLS handshake. It removes a separate client lookup to the certificate authority, saving time and improving privacy.

Reducing response payload size before transmission, commonly with gzip or Brotli for text assets. Smaller bodies shorten transfer time and lower bandwidth, especially on slow connections.

A compression algorithm that often achieves better ratios than gzip for text-based web assets. It is widely supported by modern browsers and can be applied at the edge.

Removing unnecessary characters such as whitespace and comments from source files without changing behavior. It reduces transfer size for HTML, CSS, and JavaScript.

Collecting performance data from actual visitors' browsers using APIs such as the Navigation Timing and Performance Observer interfaces. It reflects real network and device conditions rather than synthetic tests.

Performance testing using scripted, controlled requests from known locations and devices. It provides consistent baselines and alerting but may not reflect the full variety of real user conditions.

A set of user-centric metrics defined by Google covering loading, interactivity, and visual stability. They are widely used to assess real-world page experience quality.

A metric measuring when the largest visible content element finishes rendering. It is a loading-performance indicator, and faster LCP usually requires faster TTFB and efficient resource delivery.

A metric quantifying unexpected movement of visible page elements during loading. It is a visual-stability indicator and is improved by reserving space for images, ads, and embeds.

The time when the browser first renders any text or image content. It marks the start of perceived loading and is influenced by network latency, render-blocking resources, and server response time.

A resource such as a stylesheet or synchronous script that prevents the browser from painting until it is downloaded and processed. Reducing or deferring these resources speeds up first paint.

Combining multiple files, such as scripts or stylesheets, into fewer requests. It reduces request overhead but can increase cache granularity risk, so it must be balanced against caching strategy.

A security control that inspects HTTP traffic and blocks requests matching attack patterns such as SQL injection or cross-site scripting. It runs at the edge to filter malicious traffic before it reaches the origin.

Restricting how many requests a client or IP can make in a time window. It protects origins and edge resources from abuse, scraping, and traffic spikes while preserving service for legitimate users.

Detecting and controlling automated traffic, distinguishing helpful bots from abusive ones. Techniques include behavioral analysis, challenge pages, and allowlists, applied at the edge before origin contact.

Automatically redirecting traffic to a healthy alternate origin or data center when the primary becomes unavailable. It maintains availability during outages and is often paired with health checks.

Periodic probing of an origin or endpoint to confirm it is responding correctly. Results drive failover decisions and alerting, helping prevent traffic from being sent to unhealthy servers.

The disciplined process of planning, testing, approving, and deploying configuration changes with rollback plans. It reduces the risk of outages caused by untested property or DNS modifications.