Global network. Owned infrastructure.

We operate our own backbone across multiple continents with direct peering at major internet exchanges. Every server connects at 10 Gbps with DDoS scrubbing at the network edge.

10Gbps
Per server
Owned
Backbone
DDoS
Edge scrubbing
99.99%
Uptime SLA
Regions

Data center regions

Deploy in any region. Servers in the same region communicate over private networking at no additional cost.

New York
us-east-1
North America
Los Angeles
us-west-1
North America
Chicago
us-cent-1
North America
Dallas
us-south-1
North America
Toronto
ca-cent-1
North America
Amsterdam
eu-west-1
Europe
Frankfurt
eu-cent-1
Europe
London
eu-west-2
Europe
Paris
eu-west-3
Europe
Stockholm
eu-north-1
Europe
Singapore
ap-se-1
Asia Pacific
Tokyo
ap-ne-1
Asia Pacific
Sydney
ap-se-2
Asia Pacific
Mumbai
ap-south-1
Asia Pacific
Seoul
ap-ne-2
Asia Pacific
São Paulo
sa-east-1
South America

Private backbone. Direct peering. No middlemen.

Architecture

Network architecture

How we keep latency low and availability high.

Redundant uplinks

Every data center connects to at least two upstream transit providers with no single point of failure. If one provider drops, traffic reroutes automatically within seconds.

Core routers run BGP with full routing tables. We announce our own IP space and maintain direct sessions with each transit provider for maximum path control.

Network Architecture Redundant transit providers, peering at 7 IXPs, and dual-homed data centers across 28 regions TRANSIT PROVIDERS NTT (AS2914) Cogent (AS174) Lumen (AS3356) Telia (AS1299) Core Routers · BGP AS398330 · dual-stack IPv4 + IPv6 · 4.2 Tbps capacity Peering · 7 IXPs AMS-IX · DE-CIX · LINX Peering · 4 IXPs Equinix IX · NYIIX AMERICAS Virginia Oregon São Paulo Toronto EUROPE Frankfurt Amsterdam London Paris ASIA-PACIFIC Tokyo Singapore Sydney Mumbai OTHERS Johannesburg Dubai 28 regions 5 continents 4+ transit providers No single upstream dependency 7 IXPs Direct peering to major networks 4.2 Tbps Total capacity Dual-homed everywhere · automatic failover between transit providers · sub-50ms cross-region latency

DDoS scrubbing at the edge

All inbound traffic passes through our scrubbing infrastructure before reaching your server. Volumetric attacks (SYN floods, UDP amplification, DNS reflection) are filtered at line rate.

Application-layer protection is available as an add-on for HTTP/HTTPS workloads, with configurable rate limiting and WAF rules via API.

Peering

Peering and transit

We maintain direct peering at major internet exchanges to minimize hops and keep latency predictable.

Exchange / ProviderTypeCapacity
AMS-IXInternet Exchange — Amsterdam100G
DE-CIX FrankfurtInternet Exchange — Frankfurt100G
Equinix IX (NY)Internet Exchange — New York100G
LINXInternet Exchange — London100G
SGIXInternet Exchange — Singapore40G
JPNAPInternet Exchange — Tokyo40G
IX.br São PauloInternet Exchange — São Paulo40G
LumenTransit provider100G
CogentTransit provider100G
NTTTransit provider100G
Features

Network features

Included with every server at every tier.

Private Networking
Isolated VLANs between your servers in the same region. Traffic stays on our internal fabric and never touches the public internet. No bandwidth charges.
VLAN
Floating IPs
Assign IP addresses that can be moved between servers in the same region. Useful for high availability setups and zero-downtime failover.
HA
Firewalls
Stateful firewalls configurable via dashboard or API. Apply rules to individual servers or groups. Changes take effect within seconds.
Stateful
IPv4 + IPv6
Every server gets a public IPv4 address and a /64 IPv6 block. Additional IPv4 addresses available. Full reverse DNS control.
Dual stack
Load Balancers
Managed load balancers with health checks, sticky sessions, and TLS termination. Distribute traffic across servers in a region or across multiple regions.
Managed
Anycast DNS
DNS queries resolve at the nearest healthy node on our anycast network. Automatic failover during outages. Sub-millisecond resolution globally.
Anycast

See the network yourself.

Deploy a server and run your own benchmarks.