Nine points of presence, all self-operated. We do not resell other networks' capacity, which is why the map is small and the numbers are honest.
| Location | Status | Capacity | Peering |
|---|---|---|---|
| Frankfurt | Live | 400 Gbit/s | DE-CIX |
| Amsterdam | Live | 400 Gbit/s | AMS-IX |
| Stockholm | Live | 200 Gbit/s | Netnod |
| Warsaw | Live | 200 Gbit/s | PLIX |
| Vienna | Live | 100 Gbit/s | VIX |
| Bucharest | Live | 100 Gbit/s | InterLAN |
| Riga | Live | 100 Gbit/s | SMILE-IX |
| Sofia | Live | 100 Gbit/s | Bulgarian IX |
| Milan | Q3, building | 200 Gbit/s | MIX |
Open at every exchange we are present on. We do not require ratios, minimum traffic or a signed agreement — send a request from the address in your PeeringDB record and we will configure the session, usually the same week.
Private interconnects are available in Frankfurt and Amsterdam from 10 Gbit/s. We do not charge for cross-connects on our side.
All edges announce from a single origin AS. Anycast is used for the control plane and for manifest requests; media segments are served unicast from the nearest healthy edge, which keeps debugging tractable.
Health checks live at /__edge/health on every node and return a
plain 200 with the edge identifier in the body.
Announced at least 72 hours ahead, always outside 18:00–01:00 CET when viewing peaks. Edges are drained before work starts, so scheduled maintenance is invisible in delivery metrics.