Backbone
Private fiber and leased wavelengths between 41 cities. Every link has at least two physically separate paths, so a cut cable changes a route, not your uptime.
Traffic, engineered like water.
tnffx runs a private backbone between 41 cities. We read congestion the way Dutch engineers read a rising river, and move your traffic around it before your users feel a thing.
Amsterdam to Frankfurt 6.4 ms
For eight centuries the Netherlands has lived by one rule: you don't stop water, you give it a better place to go. We built tnffx on the same rule. Every packet that enters our network takes a path chosen a few milliseconds ago, from measurements taken a few milliseconds before that.
What we run
One network, operated end to end by our own engineers in Amsterdam. No resold capacity in the core, no black boxes between your users and your servers.
Edge
Points of presence inside the major internet exchanges. Your users join our network a few kilometres from where they are, not at the far end of a public transit chain.
Steering
Routes are recalculated every few seconds from live latency and packet loss. When a path degrades, traffic is already on the next best one.
Shielding
Volumetric attacks are absorbed and filtered at the edge, close to where they start. Clean traffic continues to your origin on the same low-latency paths.
How a request crosses tnffx
- 1
Arrive
The request enters the nearest edge, usually within 10 ms of the user.
- 2
Measure
Probes sent a moment earlier tell us the latency and loss on every path to your origin.
- 3
Steer
The shortest path on the map is congested, so the request takes the fastest one right now.
- 4
Deliver
It leaves our network next to your origin, and the reply comes back the same way.
Right now on the backbone
Round-trip times between our core sites, refreshed every few seconds. When a route switches to an alternate path, you see it here first.
| Route | Round trip | Path |
|---|---|---|
| Amsterdam to Frankfurt | 6.4 ms | Primary |
| Amsterdam to London | 5.8 ms | Primary |
| Amsterdam to Paris | 8.9 ms | Primary |
| Amsterdam to Stockholm | 19.7 ms | Primary |
| Frankfurt to Warsaw | 16.8 ms | Alternate, via Berlin |
| Frankfurt to Milan | 10.2 ms | Primary |
| Paris to Madrid | 17.4 ms | Primary |
| London to New York | 67.9 ms | Primary |
Bring us your hardest route.
Tell us where your users are and where your servers are. A network engineer will reply with the paths we would use and the latency you can expect on them.
Email network engineering