Networking About 5 min read

VPN Basics for Beginners: Subscriptions, Nodes, Traffic Routing, Global and Rule Modes Explained

A plain-English guide to subscriptions, nodes, route types, protocols, traffic routing, and global and rule modes—so beginners can make sense of their VPN client.

This beginner’s VPN guide starts with the client interface. A subscription provides configuration; a node is a connection endpoint you can choose; and traffic routing determines where each request goes. The difference between global and rule mode is whether the client handles all intercepted traffic the same way or evaluates it against rules. Once you understand these layers, you’ll know which settings to check when a connection fails or traffic exits through the wrong route.

How to import and update a subscription link

“Subscription” in a user dashboard may refer to a paid plan, while in a client it usually means the source of its configuration. A subscription link is the address the client uses to fetch configuration; it is not a node. It may return a node list or a configuration with policy groups and routing rules, depending on the service and the formats supported by the client. A successful import means the client read the configuration—not that a connection is active.

  1. Copy the subscription link that matches your client from the user dashboard. The dashboard URL, download links, and subscription link serve different purposes, so don’t use them interchangeably.
  2. In the client, find “Subscription” or “Configuration” and choose the option to import from a link. Check that the configuration name and node list appear, then enable the configuration.
  3. Choose a node or proxy group, then enable the system proxy or grant permission to create a tunnel as prompted. Test a service you use regularly, and check the connection log to see which rules and exit were actually used.
  4. When the service updates its configuration, refresh the existing entry instead of importing it again. If you’ve edited rules in the client, check whether refreshing will overwrite your changes and make a backup if needed.

Nodes, exit points, and route types

A node is a connection configuration you can select in the client. The entry point is where your device connects first; the exit point is where a request leaves the service network for its destination. They may be in the same location or in different ones. A node named “Singapore” may indicate the exit region, not that your device connects directly to a data center in Singapore. Check the service’s description of the node to confirm the actual route.

What route labels tell you
Label What it usually means What you can’t infer from it
Direct node Your device connects directly to the proxy node, without an additional relay entry point arranged by the service provider. Internet traffic still travels over multiple hops; this does not guarantee a shorter route or lower latency.
Relayed node Your device connects to an entry point first, then the service network forwards traffic to the exit point. A relay does not mean a dedicated line is used, nor does it tell you the bandwidth of the second leg.
IEPL dedicated line An international Ethernet private line that may carry traffic along part of the route between the entry and exit points. The route from your home to the entry point, or from the exit point to the destination, may not use a dedicated line.

Also distinguish a “direct node” from “direct” as a rule action. The former still uses a proxy node, just without an extra relay; the latter means the request does not use that client’s proxy exit. Compare routes using the same device, network, and destination service. Latency checks in a node list only show the response to a test request; they don’t predict how web pages, calls, or video will perform.

Protocols: check compatibility first

A protocol defines how the client and server communicate; it does not specify a region or route type. Nodes labeled Singapore may use different protocols, and the same protocol may be used on direct or relayed nodes. If the import succeeds but a node handshake fails, check that the protocol, transport, and security settings match, as well as the network and server status.

  • Shadowsocks: An encrypted proxy protocol. The client must support the encryption method specified in the configuration. If the configuration requires a plugin, check that it’s compatible too.
  • VMess and VLESS: These are different protocols and are not interchangeable. VLESS does not provide traffic encryption on its own, so the protocol name alone cannot tell you whether a connection is protected. Check the transport and security layers in the configuration.
  • Trojan: Typically uses TLS to protect the connection. If certificate verification fails, first check the system clock, server address, and configuration. Don’t disable verification just to dismiss the warning.
  • Hysteria2 and TUIC: Both are QUIC-based proxy protocols that rely on UDP, but their configurations are not interchangeable. If the same node works on another network but times out during the handshake on your current one, check whether that network restricts the relevant UDP traffic.

A link that works on a computer may not import on a TV. Whether a client can parse the subscription, supports the required protocols, and can handle an app’s traffic are separate questions. On Windows and macOS, the system proxy only applies to apps that use it. On iOS and Android, some clients request permission to create a system VPN; whether an app is included still depends on the system and client settings. For a TV, check whether its system can install a compatible client. For initial setup, see the device setup guides.

Choosing between global mode and rule mode

Traffic routing sends different requests along different paths. Rule mode typically matches conditions such as domain names and destination IP addresses; some clients also support matching by app or process. A matching request can connect directly, go to a specified proxy group, or be blocked. Many clients use the first matching rule, so a broad rule near the top may prevent more specific rules from taking effect. If nothing matches, check the fallback rule at the end—don’t assume it always defaults to a direct connection.

Global mode typically sends all traffic intercepted by the client through the selected outbound or proxy group. It does not automatically expand what the client intercepts, and it doesn’t mean every device, app, or DNS request uses the same exit. For example, a browser may follow the system proxy while another app uses its own network settings. Switching to global mode only changes requests already intercepted by the client.

Which mode should you use? For everyday use, start with rule mode so commonly used local services connect directly according to the rules. If a particular site has a problem, briefly switch to global mode for comparison while keeping the node unchanged. If the same app works only in global mode, first check which rule matched and which proxy group was selected rather than assuming the route is at fault.

After connecting, check traffic routing and DNS

A connected status only means that a connection has been established or proxying is enabled. It doesn’t prove that an app’s requests are using the expected exit. For example, a video app may use different domains for its home page and video content. If the home page loads, playback requests may still be matching a different rule. When reproducing the issue, review the failing and successful requests separately.

  • ✅ First check whether the client log shows the request that failed. If it doesn’t, make sure logging is enabled, then check whether the app uses the system proxy or is covered by the tunnel.
  • ✅ Check the destination domain, matched rule, and final outbound route in the log. If the request unexpectedly connects directly, see whether an earlier rule intercepted it. If it went to a proxy group, check which node the group actually selected.
  • ✅ Check the actual DNS route. The system, browser, and client may each resolve names separately; using encrypted DNS in a browser doesn’t necessarily mean queries go through the proxy. A DNS leak is when queries that should be handled by the tunnel instead go to a resolver on the local network. Intentionally configured direct DNS is different.
  • ❌ Don’t change the node, switch protocols, and clear rules all at once. Change one thing at a time, then retry the same app and action so you can tell what affected the result.

Use this order: if importing fails, check the subscription format; if the handshake fails, check the protocol and network; if only certain apps behave unexpectedly, check traffic interception, routing rules, and DNS first. For more troubleshooting, work through the troubleshooting guide. To browse available regions, see the route list.

Start a free trial