This proxy glossary defines the terms that matter when you compare a service, configure an application, or diagnose a failed connection. Each definition describes a specific technical idea without treating a proxy as a guarantee of privacy, access, or destination acceptance.
Proxy roles and access models
Proxy server
An intermediary that receives a client request, sends it toward a destination, and relays the response. Start with how a proxy server works for the complete request path.
Forward proxy
A proxy that acts for a client connecting outward to websites, APIs, or other services. It is the usual role discussed when buying private or shared proxy access.
Reverse proxy
A server-side intermediary that receives incoming traffic for one or more origin servers. It can terminate TLS, balance requests, cache responses, or protect an application.
Private proxy
A proxy address assigned exclusively to one customer under the service plan. Exclusive assignment gives a clearer reputation and load baseline, but it does not guarantee destination acceptance.
Shared proxy
A proxy address assigned to multiple customers. Sharing can lower cost, while another user may influence address load or reputation. See the private versus shared comparison.
Protocols and connection behavior
HTTP proxy
A proxy that understands HTTP requests and commonly supports HTTPS destinations through the CONNECT tunneling method.
SOCKS5 proxy
A general-purpose proxy protocol that can relay several kinds of TCP traffic and, when both client and server support it, UDP. Read the HTTP versus SOCKS5 guide.
CONNECT tunnel
An HTTP method used to ask a proxy to open a byte tunnel to a target host and port. Browsers commonly use it for HTTPS traffic through an HTTP proxy.
Proxy endpoint
The hostname or IP address and port that a client uses to reach a proxy service. Credentials or an approved source IP may also be required.
Proxy chain
A route that passes through more than one intermediary. Chains add trust dependencies, latency, and failure points, so each hop should have a defined purpose.
IP assignment and location
Static proxy
A proxy that presents the same exit IP for an extended allocation or session. It is useful when a workflow depends on a consistent network identity.
Rotating proxy
A service that changes the exit IP by request, time interval, or session rule. Rotation distributes requests but can break workflows that expect continuity.
Sticky session
A rotation option that keeps one exit IP for a defined session before changing it. The provider’s session format and duration determine the actual behavior.
Egress IP
The source IP observed by the destination after traffic leaves the proxy. It is not the only signal a destination can use to recognize a client or session.
IP geolocation
An estimate of an IP address’s country, region, or city based on commercial or public datasets. Databases can disagree or lag behind routing changes.
IP reputation
A destination or third party’s assessment of an address based on network ownership, historical behavior, complaints, observed traffic, and other signals.
Authentication and privacy boundaries
Username and password authentication
A method where the proxy validates credentials supplied by the client. Credentials should be stored and transmitted using the application’s secure configuration path.
IP allowlisting
A method where the proxy accepts connections only from approved source IP addresses. It works best when the client has a stable public IP.
DNS leak
A situation where hostname lookups use a resolver outside the intended proxy route, revealing DNS activity or producing location inconsistencies. See DNS and WebRTC leak boundaries.
WebRTC leak
A browser communication path that may expose network candidates outside an expected proxy route. Actual exposure depends on browser, network, and application behavior.
Operations and measurement
Latency
The time required for data to travel through the request path. Proxy distance, congestion, TLS negotiation, destination speed, and local network conditions all contribute.
Concurrency
The number of connections or tasks active at the same time. A service limit, application pool, or destination policy may constrain safe concurrency.
Timeout
The maximum time a client waits for connection, read, or response progress before treating an operation as failed. Different timeout stages diagnose different problems.
Rate limit
A rule that caps requests or actions during a time window. Proxies do not remove a destination’s right to enforce rate limits or acceptable-use policies.
Use the glossary as a decision checklist
Before ordering, write down the required access model, IP behavior, protocol, location, authentication method, concurrency, and destination rules. Then use the proxy selection guide to turn those terms into a controlled test plan. Current Mexela plan characteristics are summarized on the proxy pricing comparison.