How to Choose V2Ray Nodes: Latency, Traffic Multipliers, Locations, and Protocols

Compare V2Ray nodes by connection latency, traffic multipliers, exit locations, and protocol support, including VMess, VLESS, and Trojan.

At a glance

When organizing a subscription for the first time, start by ruling out nodes that don't meet your device compatibility or exit-location requirements. Retest the remaining options on the same network, then estimate data usage based on each plan's multiplier. The steps and examples here can help you choose a primary node, a separate backup, and—if needed—a low-multiplier option for large transfers.

Real connection latency: Test a complete request, not just the port

Read the millisecond figures in a node list alongside the test method. TCP latency mainly measures how long it takes your device to establish a TCP connection to the node's entry point. An open port doesn't mean authentication succeeded or that the node can reach your destination. A real connection test sends a request to a test address through the node, so the result reflects the proxy handshake, network route, and destination response. It more closely resembles an actual visit, but it's not a download-speed test and won't represent every website.

The steps below use v2rayN 7.x with the Simplified Chinese interface. Menu labels may vary between minor versions or translations. If yours differs, make sure you choose a real connection test that sends a request through the node, not a TCP probe.

  1. Check the test address

    Go to “Settings” → “Parameter Settings” and check the URL used for real connection tests. Keep the test address, device, and network the same throughout your comparison. Pause background downloads and cloud sync.

  2. Narrow down the options

    In the main node list, select three to five entries that meet your location and compatibility requirements. Right-click and choose “Test server real connection latency.” Wait for all results before starting the next round.

  3. Track variation and timeouts

    Run three rounds and record successful response times and timeouts separately. If all three succeed, use the middle result as a typical time while keeping the fastest and slowest for reference. Record any timeouts separately; don't rank them as zero milliseconds.

  4. Check real-world performance

    Right-click the node you want to test and choose “Set as active server.” Then use an application confirmed to route traffic through the proxy to open your actual destination. To compare download speeds, use the same authorized test file for each node, make sure it isn't served from a local cache, and retest during the hours you normally use the service.

These figures illustrate a selection method; they are not real-world measurements or performance guarantees. Candidate A returned 68 and 71 ms in its two successful rounds, then timed out in the third. Candidate B completed all three rounds. For frequent visits to pages with short requests, B is the better one to try first. For large downloads, compare actual throughput as well.

68 ms
Candidate A's fastest successful response
1 / 3
Candidate A's number of timeouts
96–108 ms
Candidate B's response-time range across three successful rounds

Traffic multipliers: Choose a transfer node based on data cost

A multiplier is a billing factor set by the subscription provider, not a speed boost. If the provider deducts billed traffic multiplied by the node's factor, then 8 GB of billable traffic would use 4 GB, 8 GB, or 16 GB of quota at 0.5×, 1×, or 2×, respectively. The 8 GB refers to traffic counted by the provider, which isn't necessarily the same as the file size. Check the plan details to see whether uploads and downloads are combined, whether multipliers vary by time, and when usage figures update.

Locations and routes: Choose a primary and backup based on exit requirements

A location in a subscription may describe the entry point, the exit location, or simply the provider's naming convention. Remote services usually see the exit IP address, while local latency depends on your access provider, peering routes, and congestion along the way. A short geographic distance doesn't guarantee a short network route, and two nodes in the same region can perform very differently in the evening. If your use case requires a specific location, first confirm that the target app is actually using the proxy, then cross-check the provider's exit-location details against the results you see. IP geolocation lookups can help, but databases and destination services may classify addresses differently.

How to filter by location for two common scenarios

No fixed location requirement: Start with connection stability
  • Keep one candidate from each of several regions and compare timeouts and response-time variation against the same destination.
  • Retest once during the day and once in the evening. If the ranking reverses, favor the node that performs better during your usual usage hours—not one that's fast only when the network is quiet.
Fixed location requirement: Verify the actual exit location first
  • Rule out nodes whose exit locations don't meet your requirements, then compare connection performance and multipliers among the remaining options.
  • If the provider can explain how routes differ, favor a backup with a different entry point or upstream route. Different suffixes in node names don't guarantee independent failure paths.

The exit location determines whether a node fits your use case; the access route determines the experience. Verify both separately, and compare only nodes that pass both checks.

VMess, VLESS, and Trojan: Check the core and transport combination first

A protocol name describes how a connection works; it doesn't predict performance on its own. Two nodes labeled VLESS may use different transports and security layers, and may connect through different entry points, exit locations, or load conditions. Seeing a node after importing a subscription only means the client recognized the entry. Whether it works also depends on whether the active core can handle the complete configuration.

VMess + WebSocket + TLS

Proxy protocol
VMess
Transport
WebSocket
Security layer
TLS
Fields to check
User ID, path, TLS server name

If the path or TLS server name doesn't match the server configuration, the connection may fail even when the port is reachable.

VLESS + REALITY

Proxy protocol
VLESS
Example transport
TCP
Security layer
REALITY
Fields to check
User ID, server name, public key, short ID

The core must support this configuration. If the node specifies a flow-control setting, it must also match the server.

In common configurations, Trojan uses password authentication and relies on TLS. If certificate validation fails, check your system clock, the server name supplied by the subscription, and whether they match the server certificate. Don't disable certificate validation just to get a test to pass. The two cards above are examples of configuration combinations, not templates for manually converting an existing node to a different transport.

Make your final picks: Primary, backup, and high-traffic nodes

Node rankings don't need to be permanent. Your primary should reliably handle real tasks on your usual devices and during your usual hours. Your backup should take over the same tasks if the primary fails. Keep a low-multiplier node only if you actually need to transfer large amounts of data. Results on broadband, mobile networks, and different Wi-Fi access points aren't directly comparable. After switching networks, retest your two or three shortlisted nodes instead of scanning the entire subscription again.

Download v2rayN