Nodes are placed in carrier-neutral facilities with direct access to regional exchange points. Each PoP can absorb the traffic of its nearest neighbour during maintenance.
| Location | Facility | Exchange | Capacity | Status |
|---|---|---|---|---|
| Vilnius, LT | Vilnius North | LIXP | 2 × 25 Gbit | live |
| Riga, LV | Riga Central | LIX | 2 × 10 Gbit | live |
| Warsaw, PL | Warsaw Wola | PLIX, THINX | 2 × 25 Gbit | live |
| Frankfurt, DE | Frankfurt East | DE-CIX | 4 × 25 Gbit | live |
| Amsterdam, NL | Amsterdam Zuidoost | AMS-IX, NL-ix | 2 × 25 Gbit | live |
| Stockholm, SE | Stockholm Kista | Netnod, STHIX | 2 × 10 Gbit | live |
We peer openly at every exchange where we are present. No minimum traffic level, no ratio requirement, no fee. Sessions are usually configured within two business days.
A registered ASN, up-to-date entries in a recognised routing registry, a maintained NOC contact and consistent announcements across all shared exchanges.
Prefixes are validated against RPKI and IRR data. Invalid announcements are dropped rather than depreferenced. Maximum prefix limits are set per session and reviewed quarterly.
Cross-connects are available in all listed facilities for partners exchanging sustained traffic above 5 Gbit. Contact the NOC for a port assignment.
Every PoP carries at least two transit providers plus local exchange capacity, so a single upstream withdrawal never isolates a region.
Intra-Baltic under 12 ms, Baltic to Frankfurt under 30 ms, Baltic to Amsterdam under 34 ms, measured at the 95th percentile.
Withdrawal of a PoP shifts its traffic to the neighbour with the lowest measured path cost. Convergence is typically under 40 seconds.
Continuous active probing between all PoPs and against a fixed set of external reference targets, sampled every 30 seconds.