V2Ray Glossary

Why can a node pass a latency test while websites still won’t load? What’s the difference between system proxy and TUN mode? Find the term in the client, learn what it does, then decide whether to change the setting.

Protocols & Transport

A protocol defines how two endpoints exchange proxy traffic; transport and security settings are separate.

VMess
VMess is a protocol clients and servers use to exchange proxy traffic. If you see VMess on a node’s edit page, that node communicates with its remote endpoint using the VMess format. The client also needs details such as a user ID that matches the server. Two protocols can use the same address and port, but you can’t freely swap one for the other in the client.
VLESS
VLESS is a proxy protocol that does not encrypt traffic on its own. If a node’s details list VLESS, TCP, and TLS, each refers to a different setting: the proxy protocol, the transport, and transport security. When troubleshooting a connection, check each one separately—matching protocol names alone doesn’t mean the configuration is correct.
Trojan
Trojan is a protocol that sends proxy requests over a TLS connection. When it appears in a node’s Type field, it identifies the protocol; it isn’t a security warning from the client. Trojan typically uses a password for authentication. If the log reports a certificate error, the issue may be with TLS verification rather than an incorrect password.
REALITY
REALITY is a transport security mechanism supported by Xray, not a proxy protocol alongside VLESS. A single node configuration can therefore specify both VLESS and REALITY in separate fields. If the client doesn’t show the corresponding option after import, first check whether the core in use supports that combination. Don’t change the security type to something else on your own.

Cores & Ecosystem

The interface lets you select nodes and change settings; the core is what actually handles connections.

Project V
Project V is the name of a project in the V2Ray ecosystem, not an additional component to select during installation. It began as an effort to build network proxy tools, with V2Ray as one of its core tools. When a tutorial mentions Project V, the core, and a graphical client, it’s usually referring to three different layers: the ecosystem, the runtime component, and the user interface.
Core
The core is the program that actually establishes connections and handles traffic behind the client. The graphical interface saves nodes, organizes settings, and starts the core; seeing the window open doesn’t mean the core started successfully. If nothing connects after you import a node, check the startup log for errors, then work out whether the core failed to run or the node details are incorrect.
Xray
Xray is a proxy core forked from the V2Ray project that later developed independently. It shares some protocol and configuration concepts with the V2Ray core, but features developed independently aren’t necessarily interchangeable. For example, v2rayNG uses the Xray core. When checking feature support, look at the core actually running rather than relying on v2ray in the client’s name.
V2Fly
V2Fly is a community project that continues to maintain and develop the V2Ray core. You can think of v2flyNG as a graphical client that uses the V2Fly core, not as the name of a protocol. In a tutorial, the client name helps you find the right interface, while the core name tells you whether a configuration field is supported.

Subscriptions & Nodes

A subscription imports configurations; a node carries your requests. They’re separate things and should be troubleshot separately.

Subscription
A subscription URL is an address a client uses to retrieve a set of node configurations. After saving the URL to a subscription group, you still need to select Update Subscription before the client fetches the list. A successful update doesn’t mean you’ve switched to a node. Subscription URLs may contain access credentials, so hide the full link before sharing a screenshot for help.
Node
A node is a connection configuration you can select from the client’s list. It usually includes a server address, port, protocol, and authentication details. A prominent name in the list is often just a label and doesn’t necessarily identify a real data center. Selecting a node sets the route to use; whether an app uses it also depends on how its traffic reaches the client and on the routing rules.
Subscription group
Subscription groups keep nodes from different subscriptions organized separately. In v2rayN, you can update the current group from the Subscription Groups menu instead of refreshing every source each time. Groups control where list entries come from and which subscriptions are updated; they aren’t routing groups. Switching groups to view another list won’t automatically reroute traffic from all your apps.
Subscription conversion
Subscription conversion reformats node details into a format another client can read. It doesn’t turn one protocol into another, and fields unsupported by the target format may be dropped even when conversion succeeds. If the original provider offers a compatible format, use that instead. This avoids handing a subscription URL containing credentials to an unfamiliar service.
TCP latency
TCP latency usually means the time it takes to establish a TCP connection to a node’s address and port. This test only tells you whether the port is reachable and how long the connection takes. It doesn’t verify the proxy protocol handshake or try to load a webpage. So even if the TCP test succeeds, check the protocol, authentication, and transport settings if requests still fail.
Real connection latency
Real connection latency measures how long a request takes to reach a test URL through a node. It’s closer to everyday use than a port-only test, but results also depend on the test site, your current network, and the client implementation. A timeout doesn’t necessarily mean the node is broken, and a successful test doesn’t guarantee fast downloads. For meaningful comparisons, use the same test URL.

Routing & DNS

Routing determines where requests go; DNS resolves domain names to addresses. Web requests and DNS lookups don’t necessarily take the same route.

Split tunneling
Split tunneling sends different requests through different routes. For example, you can connect directly to your home router’s admin page while routing requests that match proxy rules through the selected node. Split tunneling only applies to traffic that reaches the core. If an app isn’t using the system proxy and its traffic isn’t captured by TUN, changing the rules won’t change its route.
Routing rule
A routing rule determines which traffic goes through which route. A rule usually combines a match condition with an action, such as sending traffic matching an address group directly. The core typically checks rules in order and uses the route for the first match. If an earlier rule has a broad match, a more specific rule later in the list may never be reached.
GeoIP
GeoIP rules match a set of IP address ranges. For example, geoip:private can match common private or local address ranges; the exact coverage depends on the rule data. GeoIP checks the destination address, not the website name. For requests made to a domain, whether the IP is resolved before matching depends on the current routing resolution strategy.
GeoSite
GeoSite rules match a set of domain names. For example, geosite:cn refers to a domain group named cn in the rule database; it doesn’t check where a server is physically located. Whether a particular domain is included depends on the database used by the client. The same rule syntax can produce different matches with a different set of rule data.
FakeDNS
FakeDNS maps a domain to a temporary address so the core can identify subsequent connections. When an app connects to this address, the core can use the mapping to recover the original domain and apply domain-based routing. This address is used internally by the client to associate requests; it isn’t the website’s actual server address. Whether FakeDNS works correctly depends on subsequent connections also being handled by the core.
DNS leak
A DNS leak occurs when domain lookups use an unexpected resolver path. For example, webpage traffic may go through a proxy while the app sends domain lookups to the resolver currently used by the system. To investigate, first determine which lookups are supposed to go direct, then check which resolver is actually being used. The webpage’s apparent exit IP alone doesn’t prove that DNS took the same route.

Client Settings

First identify whether a setting controls how apps connect, how traffic is transported, or which port the client listens on.

System proxy
The system proxy tells apps that follow system settings where to find a proxy. In the v2rayN Windows interface, for example, Auto Configure System Proxy in the tray menu sets the proxy settings, while Clear System Proxy removes them. Apps with their own proxy options may ignore the system settings. This setting also doesn’t route all traffic from the device.
TUN mode
TUN mode receives traffic routed to it through a virtual network interface. If you see Enable Tun in the v2rayN main window, think of it as another way for traffic to reach the client: apps don’t need to read the system proxy settings, and traffic matching the routing rules can still be handled by the core. The rules still determine whether traffic uses a proxy; reaching the core doesn’t mean everything is forwarded to a node.
Mux multiplexing
Mux multiplexing carries multiple logical connections over fewer underlying connections. It may reduce the overhead of repeatedly establishing connections when opening many short-lived connections, but it isn’t a bandwidth-boost switch. Whether it works depends on the protocol and both endpoints’ implementations. To troubleshoot instability, keep other settings unchanged and compare performance with Mux on and off.
Local listening port
A local listening port is where an app sends requests to the client running on the same device. When configuring a proxy manually in an app, check the address, port, and interface type the client is actually listening on rather than copying the remote port from the node list. Even with the right port number, a connection may fail if you configure an HTTP proxy as SOCKS.