Proprietary Ping Network
Real-time multiplayer server discovery
How It Works
Distributed Latency Measurement
ServerPulse operates a global mesh of over 340 probe nodes spanning 48 countries, each continuously measuring round-trip latency to thousands of game servers. Unlike single-point ping checkers, our network triangulates your true connection quality by comparing responses across the nearest up to six probes relative to your IP geolocation.
Every 90 seconds, each probe runs parallel ICMP and TCP SYN tests against registered game server endpoints on standard ports — UDP 27015 for Source engines, TCP 7777 for Rust, UDP 2346 for CS2 matchmaking, and custom ports for mods and private lobbies. Results are aggregated, outlier-filtered using a median-of-medians algorithm, and pushed to the browser UI in under 400ms. The entire pipeline runs on a Kafka-backed stream with Redis caching, ensuring that the ping value you see is never older than 120 seconds.
Why It Matters
Built for Competitive Accuracy
One ping from your ISP's DNS resolver tells you almost nothing about your actual gameplay experience. Our multi-node approach gives you the truth.
Multi-Node Triangulation
Latency is measured from the three nearest probes to your IP, then averaged. If you're in Frankfurt, probes in Frankfurt, Amsterdam, and Zürich cross-verify the result — eliminating single-hub skew.
Protocol-Aware Testing
We don't just ping an IP. We send protocol-specific handshake packets for each game engine — GoldSrc, Source 2, Unreal, Unity — so the reported latency reflects the actual traffic your client will send.
Jitter & Packet-Loss Metrics
Beyond raw ms, each probe reports jitter variance and packet-loss percentage over the last 300 samples. High jitter on a low-ping server often means a congested upstream hop — we flag it so you can avoid rubber-banding.
Historical Baselines
Every probe retains 30 days of per-server latency data. You can see whether a server's ping has degraded over the past week, helping you distinguish a bad day from a permanently poor route.
AS-Level Routing Intelligence
Our probes tag each result with the Autonomous System path. If your connection to a Dallas server routes through a congested European backbone, you'll see the AS hop count and can choose a server with a cleaner path, even if its raw ping looks similar.
Open Probe API
Third-party developers can query our probe results via REST or WebSocket endpoints. Over 120 community tools — from Discord bots to in-game overlays — already integrate ServerPulse latency data.
Common Questions
Frequently Asked Questions
How many probe locations do you operate?
As of Q2 2025, we run 342 active probes across 48 countries. Major clusters exist in Ashburn VA (48 probes), Frankfurt DE (36), Tokyo JP (28), São Paulo BR (18), and Singapore SG (22). We add new probes monthly based on community requests and server-density heat maps.
How often are ping measurements updated?
Each probe tests every tracked server every 90 seconds. Aggregated results are refreshed in the browser UI every 15 seconds. If a probe goes offline, its weight is removed from the average immediately and the next-nearest probe fills the gap.
Why does my in-game ping differ from ServerPulse?
Game clients measure latency from the moment your packet leaves your machine to the moment the server acknowledges it — including your router, ISP, and the server's own load. We measure the network path only. If a game server is CPU-bound, its in-game ping will be higher than our network-only number. That's why we also report server tick-rate and player count alongside latency.
Can I add my own server to the ping network?
Yes. Register your server through the dashboard at servers.serverpulse.io/register. Once approved, all probes begin testing your endpoint within 5 minutes. You'll receive a webhook notification when your first latency readings are live, and you can view real-time probe results from any location on our map.
Do you support non-standard or modded game ports?
Absolutely. Our probe fleet tests any TCP or UDP port. Modded communities — from MTA:SA on port 22003 to Garry's Mod DarkRP servers on 27016 — are fully supported. When you register a server, simply specify the protocol and port, and our probes adapt automatically.
Is the probe data available for research or public use?
We publish anonymized aggregate latency datasets quarterly on our open-data portal. Individual probe results are accessible via API key for developers and researchers. Contact data@serverpulse.io for bulk access or academic partnerships.