Understand common proxy terms in a real connection flow
A term becomes useful when it explains the request path, scheduling rules, or capacity limits. This glossary gives practical definitions and the boundaries to verify in product documentation.
How the terms fit together
A proxy request normally starts with authentication, enters the proxy network through a supported protocol, and uses an exit IP selected from an IP pool. A session links related requests, rotation determines when the exit changes, and concurrency limits how much work happens at the same time.
“Residential,” “SOCKS5,” “allowlist,” and “cleanliness” describe address origin, transport, access control, and address quality. They are not competing values on one scale. When reading specifications, ask how each term is defined and measured by that particular service.
Addresses and network identity
Residential describes address origin; exit IP describes where traffic leaves the proxy network.
Residential IP
A residential IP generally means a public address assigned through an internet service provider and associated with household broadband or a consumer ISP network. Many product systems classify mobile-network addresses separately as mobile proxies, so check the provider's exact boundary. The term describes network origin. It does not mean the address belongs to a specific individual, nor does it prove a request's purpose, anonymity, or outcome.
Exit IP
The exit IP is the source address used when the proxy connects to the destination on the client's behalf. The destination normally uses it to infer network ownership and approximate region. The proxy entry hostname you configure may not be the final exit, so record the observed exit with a test request.
Pools, sessions, and rotation
See how an address collection, continuity, and switching rules work together.
IP pool
An IP pool is a collection of exit addresses available to the proxy scheduler. Pools may be divided by country, region, network, product type, or online status. A published global total is not the number of usable addresses in one region at one moment.
Session
A session is the connection or scheduling context that associates a set of requests. A service may identify it through a username parameter, session ID, port, or connection lifetime. A session does not automatically preserve one IP; verify sticky-session support, maximum duration, and invalidation conditions.
Rotation
Rotation is the rule for changing an exit IP. Common triggers include every request, every new connection, session expiry, an explicit parameter change, or an unavailable address. Compare not only when an IP changes, but also whether it can stay stable and what happens after failure.
Allowlists and SOCKS5
One controls who may connect; the other defines how traffic is transported.
IP allowlist
The service accepts proxy connections only from public source IPs registered in advance. It suits servers with a stable egress address; a changed local or cloud public IP normally requires an allowlist update.
Register the public source IP used to reach the proxy, not the proxy exit IP you want to receive.
SOCKS5
SOCKS5 is a lower-level proxy protocol that can relay several kinds of TCP application traffic and may support UDP and authentication. Exact capabilities depend on both client and server implementations.
Confirm native client support and verify DNS, UDP, and authentication behavior for the workload.
Username-and-password authentication
The client supplies credentials when connecting, and the service uses them to determine account access. This is portable across source networks, but leaked credentials may be abused.
Use secure secret management and never place proxy credentials in a public repository or client-side code.
Concurrency and IP cleanliness
Separate capacity limits from address-quality descriptions and outcome claims.
Concurrency
Concurrency is the number of connections, requests, threads, or tasks active at the same time. Services may count different units, and connection count is not the same as requests per second. At the limit, new work may be rejected, queued, rate-limited, or time out, so verify the exact definition first.
IP cleanliness
Cleanliness is not a single industry standard. A service may base it on address history, degree of sharing, risk-database records, or an internal score. A higher advertised score cannot guarantee acceptance by a destination. Treat it as supporting information and validate performance with compliant, representative requests.
How to read product specifications
Turn glossary terms into configuration details that can be compared and tested.
- 01
Confirm the unit of measurement
Separate traffic, IP count, ports, connections, requests, and duration before comparing plans.
- 02
Confirm exit behavior
Record region selection, session duration, rotation triggers, and handling of an unavailable address.
- 03
Check connection requirements
Verify supported protocols, authentication, DNS behavior, entry endpoints, and client compatibility.
- 04
Validate limits and outcomes
Use a representative workload to observe concurrency limits, error types, latency, retries, and actual success rates.